CN1162001C - Motion picture coding apparatus and method for coding a plurality of moving pictures - Google Patents

Abstract

The moving image device for encoding a plurality of moving image signals of the present invention includes: a plurality of encoding sections for encoding a plurality of moving image signals, and the encoding section provides quantization of a plurality of moving images with a preset quantization width. a plurality of quantized sub-sections of the image signal; and a quantization control section for controlling the quantization width simultaneously used by the quantization section.

Description

Motion picture coding apparatus and method for coding a plurality of moving pictures

technical field

The present invention relates to a moving picture encoding apparatus/method for encoding a plurality of moving pictures in such a manner that between compressing/encoding a plurality of moving pictures and multiplexing the encoded signals to The quality of the motion picture reproduced by a decoding device after the motion picture coded signal is generated at a fixed bit rate is improved.

The present invention also relates to a moving picture for multiplexing a moving picture encoding signal obtained after a process of compressing/encoding a plurality of moving pictures or a moving picture segment obtained by separating a high-precision moving picture Like multiplexing means/methods.

The invention further relates to an image transmission device which compresses/encodes a plurality of moving images and transmits the resulting encoded moving image signals at a limited bit rate for use in satellite communications or network communications such as a LAN. Motion picture transmission.

Background technique

In one of these methods for compressing/encoding and multiplexing a plurality of moving pictures to transmit the resulting moving pictures as a bit stream at a fixed bit rate, a preset The fixed bit rate so that the sum of the number of bits representing a plurality of moving pictures after compression/encoding constitutes a fixed bit rate. When an image of high complexity is input in such a method, the specified bit rate is not appropriate, with the result that the image quality is greatly degraded. Conversely, when inputting low-complexity images, stuffing increases, resulting in wasted use of the bit rate. As a result, two opposing problems arise. That is, when trying to use the bit rate efficiently, the picture quality degrades. Conversely, when trying to make the picture high quality, the bit rate cannot be effectively used, reducing the number of moving pictures that can be multiplexed.

In order to solve the above problems, Japanese Patent Application JP8-23540 discloses an encoding control device in which the bit rate at the time of encoding of each moving image signal can be changed to improve the quality of moving images reproduced by a decoding means. The moving picture reproduced from the moving picture coded signal is compared with the original picture in this encoding control device. The bit rate is increased when a reproduced moving picture is more distorted than the original picture, and the bit rate is decreased when a reproduced moving picture is less distorted than the original picture so that the The quality of moving pictures becomes uniform and thus improves the quality of the entire picture.

However, according to the method described above, the bit rate is changed after the amount of distortion is determined. Therefore, in the case where the moving image suddenly becomes complicated due to a scene change or the like, the quality of the moving image becomes lower immediately after the scene change.

In addition, in such a conventional apparatus in which the bit rate is variable when encoding a moving picture, when a variation such as the bit rate is proportional to the number of bits generated within a predetermined time interval, a simple multiplexing operation is performed. The input buffer of the decoding device may cause overflow or underflow during implementation.

In Japanese Patent Application JP-6-6777 there is disclosed an encoding apparatus in which encoded motion pictures have different resolutions. Fig. 37 has shown the composition of this encoding device, divides high-quality HD speaker/moving image into several blocks by a frame synthesis block part 1101, and sends these blocks that are divided into a distribution part 11020 as a signal switch When 1107 is switched to an end connected to the block allocation section 1102 in response to an operation mode signal operation, the block allocation section 1102 distributes each block received from the frame synthesis block section 1101 to the moving image coding section 1103 to 11060 The moving image encoding sections 1103 to 1106 encode the respective blocks and output encoded resultant signals. When the signal switch 1107 is switched to the other end, the h/moving picture with a normal quality is supplied to the moving picture coding section 1103 through the signal switch 1107. The moving image encoding section 1103 encodes the R/moving image and outputs the resulting encoded signal.

Therefore, in this conventional apparatus, a low-resolution moving picture signal and a high-resolution moving picture signal can be encoded by switching the signal switch 1107.

However, the conventional apparatus having the above constitution does not give the concept of simultaneously encoding and multiplexing moving pictures with different resolutions. Therefore, this device is not suitable for simultaneous transmission of moving pictures with different resolutions used in broadcasting and the like.

In a conventional technique of multiplexing a plurality of moving pictures for transmission, a fixed bit rate is specified for each encoding device and the moving pictures are encoded at the specified bit rate by each encoding device and multiple The resulting motion picture coded signal is multiplexed. However, this technique has the following problems. The complexity of motion pictures varies over time. Therefore, it is necessary to set each bit rate in order to make the degradation of the image quality of the encoded and transmitted moving images imperceptible, so that even when the most complicated moving images are encoded, the reproduced moving images The degradation of image quality is also imperceptible. This makes the transfer less efficient.

In order to solve the above-mentioned problems, a method of varying the bit rate in accordance with the complexity of moving image coding has been proposed. According to this method for simple motion pictures, the bit rate is reduced to obtain efficient transmission while multi-channel communication using multiple channels can be the same transmission bit rate. Such a method is disclosed, for example, in WO95/32565.

Fig. 38 shows an apparatus for carrying out the method described above. Referring to FIG. 38, moving picture encoding sections 2101, 2102, and 2103 encode input moving pictures according to respective target rates specified by a multiplexing section 2104. And output the resulting coded signal to the multiplexing section 2104. The multiplexing section 2104 multiplexes the received coded signal and outputs the multiplexed signal to a transmission line. The multiplexing section 2104 also adjusts the target rate according to the transmission rate used for encoding the signal and supplies the new target rate to the moving picture encoding sections 2101 to 2103.

The target rate is set such that the sum of the target rates is equal to the transfer rate. In more detail, each of the moving picture encoding sections 2101 to 2103 transmits the amount of encoding distortion at the time of encoding to the multiplexing section 21040 every fixed period. The target rate is adjusted by the amount of coding distortion transmitted by each of the image coding sections 2101 to 2103, and the sum of the target rates is equal to the transmission rate. That is, the target rate is increased for a moving image encoding section that transmits a large amount of distortion, and is decreased for a moving image encoding section that transmits a small amount of distortion. In this manner, moving pictures encoded by the moving picture encoding sections 2101 to 2103 are provided with a designated bit rate according to the complexity of the encoding.

However, in the conventional method described above, the target rate is corrected so that the amount of distortion is reduced after the distortion is generated. Therefore, image quality tends to decrease immediately after, for example, a scene change to a scene that is difficult for encoding. Also, in this conventional method, the multiplexing section calculates a target rate to be allocated to the moving image encoding section based on the amount of encoding distortion in the moving image encoding section. Therefore, when the number of moving picture coding sections connected to the multiplexing section is large, the amount of calculation required to specify the target rate increases. In addition, if these moving image encoding sections are dispersed in a network and moving images are transmitted in the network, processing of information collected from encoding distortion of each moving image encoding section and designation of each moving image encoding section are required. target rate. It is difficult to realize these real-time processing.

Contents of the invention

A moving picture encoding apparatus for encoding a plurality of moving picture signals of the present invention includes: a plurality of encoding sections for encoding a plurality of moving picture signals. a plurality of quantization subsections of a moving image signal; and a quantization control section for simultaneously controlling the quantization width used by the quantization subsections.

With the above constitution, the quality of the output moving image can be made uniform. In one embodiment of the invention the quantization widths used by the quantization subsections are the same.

In another embodiment of the present invention, the moving image coding apparatus further includes a plurality of accumulating sections for temporarily accumulating a plurality of moving image encoded signals output from a plurality of encoding sections, wherein the quantization control section is based on the The sum of the number of bits of a plurality of moving image coding signals accumulated by the accumulation section determines the quantization width used by the quantization subsection.

In still another embodiment of the present invention, the moving picture coding apparatus further includes a screen separation section arranged upstream of the plurality of coding sections for separating an input moving picture signal into a plurality of moving picture signals.

According to another aspect of the present invention, there is provided a moving picture multiplexing apparatus for multiplexing a plurality of moving picture encoded signals obtained by encoding a plurality of moving picture signals. The apparatus includes: a plurality of accumulation sections for temporarily accumulating a plurality of moving image coded signals; and for multiplexing the plurality of moving image coding signals accumulated in the plurality of accumulation sections and performing the operation at a fixed bit rate. Outputting the multiplexed portion of the multiplexed signal; used to calculate a specific bit from each of a plurality of moving image coded signals is input to a decoding device until the specific bit is decoded by the decoding device The decoding delay time calculation part of one time period until the code is coded; the frame number of bits for calculating the number of bits of one frame of a specific bit of each of a plurality of moving image coding signals accumulated in a plurality of accumulation parts Calculating section; and being used for controlling the number of bits of a plurality of moving image coded signals input from a plurality of accumulating sections to the multiplexing section according to the calculation result of the decoding delay time calculating section and the frame bit calculating section The multiplexing control part of the first input bit.

With the above constitution, multiplexing is carried out in consideration of the decoding delay time in the decoding means. Therefore, the occurrence of overflow and underflow can be avoided when decoding a moving image coded signal by the decoding apparatus.

In one embodiment, the multiplexing control part includes: a first parameter generation subsection, used to determine a plurality of first parameters, wherein each parameter in the plurality of first parameters is determined by the frame number calculation part The ratio of the calculated bit number to the time calculated by the decoding delay time calculation part; the second parameter generation subsection is used to determine the second parameter, which is the sum of a plurality of first parameters; the third parameter generating a subsection for determining a plurality of third parameters, wherein each of the plurality of third parameters is one of the plurality of first parameters for a plurality of moving image encoded signals accumulated in a plurality of accumulation sections a ratio of each to said second parameter; and a multiplexing bit calculation section for calculating a plurality of first input bit numbers according to a plurality of third parameters.

In one embodiment of the present invention, the moving image multiplexing device further includes: an accumulation bit calculation section for calculating the number of bits of the plurality of moving image encoding signals accumulated in the plurality of accumulation sections; An unoccupied capacity calculation section of calculating an unoccupied capacity of an input buffer of the decoding means immediately before specific bits of each of a plurality of moving image coded signals are input to the decoding means, wherein the multiplexed number of bits The calculation section sets, for each of the plurality of second input bit numbers, at (1) the first input bit number determined based on the third parameter (2) the bit number of the moving image coded signal calculated by the accumulated number of bit number calculation section. and (3) a minimum value among the unoccupied capacities of the input buffers of the decoding device calculated by the unoccupied capacity calculation section, and the multiplexing bit calculation section controls The number of bits is the number of bits of the motion picture coded signal sent from the accumulating section to the multiplexing section.

In another embodiment of the present invention, when the plurality of second input bit numbers calculated by the multiplex bit number calculation section is different from the corresponding first input bit number, the multiplex bit number calculation section Increment the second input bit.

According to still another aspect of the present invention, there is provided a moving picture coding method for simultaneously coding a plurality of moving picture signals. The method includes encoding all of the plurality of moving image signals with the same quantization width.

In addition, the moving picture coding method of the present invention for simultaneously coding a plurality of moving picture signals includes using the sum of the bits of a plurality of moving picture encoding signals generated by encoding a plurality of moving picture signals and the temporal The quantization width obtained by accumulating the difference between the output bits of the plurality of encoded moving image signals to be outputted encodes the plurality of moving image signals.

According to still another aspect of the present invention, there is provided a moving picture multiplexing method for multiplexing a plurality of moving picture coded signals after being accumulated by a plurality of accumulation sections. The method comprises the steps of: (1) calculating the time period from when each specific bit of a plurality of moving image coded signals is input to the decoding device until the specific bit is decoded; (2) calculating the time period included in the multiple The number of digits of a frame of a specific bit of each of a plurality of moving image coded signals accumulated by the accumulation parts; (3) from the plurality of moving image coded signals according to the calculation results from steps (1) and (2) The number of multiplexing bits to be multiplexed is determined in .

In one embodiment of the present invention, step (3) comprises: (4) determine a plurality of first parameters, each of a plurality of first parameters is the number of digits obtained in step (2) and the number obtained in step (1) (5) determine a second parameter that is the sum of a plurality of first parameters; (6) determine a plurality of third parameters, each of a plurality of third parameters is a plurality of first parameters A ratio of each to the second parameter for the plurality of moving image coded signals; and (7) determining a plurality of multiplexing bits based on the plurality of third parameters.

In yet another embodiment of the present invention, the number of input multiplexing bits for multiplexing each of a plurality of motion picture coded signals is set to be multiplexed at (1) determined according to the third parameter The minimum value among the number of digits of the moving image coded signal accumulated in the accumulation part and (3) an unoccupied capacity of the input buffer of the decoding device by the multiplexing number of bits (2), and is multiplexed Each of the multiplexed bits is controlled using the input multiplexed bit.

In yet another embodiment of the present invention, when at least one input multiplexing bit is different from the corresponding multiplexing bit, the input multiplexing bit is incremented.

Also, in the moving picture multiplexing method for multiplexing a plurality of moving picture coded signals of the present invention, a plurality of multiplexing bit numbers corresponding to a plurality of moving picture coded signals are Preset to zero. The method includes the steps of: (1) calculating the number of frame bits of a frame comprising a specific bit of each of a plurality of moving image coding signals; (2) calculating until each of the plurality of moving image coding signals A time period until it is decoded; (3) select the moving image coded signal whose time period obtained in step (2) is shorter than a preset time value; (4) according to the time period selected in step (3) (5) calculate the sum of multiple main multiplexing bits; (6) combine multiple main multiplexing and (7) update the preset time value, repeating steps (1) to (7), and if the plurality of main multiplexing bits is greater than one The number of bits set, a plurality of sub-multiplexing bits are determined and added to the plurality of multiplexing bits to multiplex the motion picture coded signal of the plurality of multiplexing bits.

Therefore, since the encoded moving image signal is multiplexed earlier than the encoded signal of the decoded frame, the encoding of the moving image by the decoding means The decoding of the signal prevents the occurrence of overflow or underflow.

In addition, the moving picture multiplexing method for multiplexing a plurality of moving picture coded signals in the present invention includes the step of: The number of bits of the frame determines the number of minimum transfers required to prevent underflow of a plurality of input buffers of the decoding device and determines the number of transfers required until the frames of the plurality of motion image coded signals are decoded; determining a plurality of lower limits of the number of transfer bits based on the plurality of minimum transfer numbers; and multiplexing the plurality of lower limits of the number of transfer bits before multiplexing the plurality of moving image coded signals.

In addition, the moving picture coding apparatus for coding a plurality of moving picture signals of the present invention includes: a plurality of coding sections for coding a plurality of moving picture signals, each of which includes a A preset parameter quantizes the quantization sub-section of the moving picture signal and is used to stuff the stuffing sub-section of the quantized moving picture signal with a preset stuffing amount; A quantization control section of parameters used partially simultaneously; a plurality of accumulation sections for temporarily accumulating a plurality of moving image encoded signals output from the plurality of encoding sections; for multiplexing in the plurality of accumulation sections a multiplexing section that accumulates a plurality of moving image coded signals and outputs a multiplexed signal at a fixed bit rate; a stuffing control section for controlling a stuffing amount used by the stuffing section; and for calculating an unoccupied capacity of each of the plurality of input buffers of the decoding device after a predetermined time period when the plurality of moving image coded signals are not input to the input buffer for the predetermined time period A virtual unoccupied capacity calculation section, wherein the quantization control section determines the parameter used by the quantization subsection according to the sum of the number of bits of a plurality of moving image coding signals, and the stuffing control section determines the parameter used by the quantization subsection according to the sum of the bits in the plurality of accumulation sections The sum of the number of bits of the moving image coding signal accumulated in the fixed bit rate, and the number of bits of each moving image coding signal accumulated in the accumulation section and the input obtained by the virtual unoccupied capacity calculation section The difference between the unoccupied capacity of the buffer to control each stuffing portion.

In addition, the moving picture encoding method of the present invention is based on the sum of the bits of the plurality of moving picture encoding signals generated by encoding a plurality of moving picture signals and the sum of the bits of each of the plurality of moving image encoding signals after temporary accumulation. A difference between an output number of bits output is used to simultaneously encode a plurality of moving image signals using each preset quantization width, wherein occupancy of a plurality of encoding buffers of the moving image encoding signals temporarily accumulated is calculated , calculating the unoccupied capacity of each of a plurality of input buffers of the decoding device when the motion picture coded signal is not input to the input buffer for the preset time period after a preset time period; Subtracting the unoccupied capacity of the input buffer from the coded buffer occupancy of a preset value to obtain a buffer difference for each moving image coded signal, by applying a value larger than the preset value that does not affect the coded distribution Stuffing is performed by adding a buffer difference value of a set value to a one-digit signal obtained from a motion picture encoded signal in which the buffer difference value is smaller than the preset value.

In one embodiment, the preset value is 0.

In addition, the moving picture encoding apparatus for encoding a plurality of moving picture signals of the present invention includes: a plurality of encoding sections for encoding a plurality of moving image signals, each of the plurality of encoding sections includes a Preset quantization width quantization of the quantization subsection of the moving image signal; a quantization control section for controlling parameters simultaneously used by the quantization subsection; used for temporarily accumulating a plurality of moving images outputted from a plurality of encoding sections A plurality of accumulation sections of coded signals; a multiplexing section for multiplexing a plurality of motion image coded signals accumulated in the plurality of accumulation sections and outputting the multiplexed signal at a fixed bit rate ; One is used to calculate the non-occupancy of each of a plurality of input buffers of the decoding device when a plurality of moving image coded signals are not input into the input buffer for the preset time period after a predetermined period of time A virtual unoccupied capacity calculation section of capacity, wherein the quantization control section is based on the sum of the number of bits of the moving picture coded signal accumulated in a plurality of accumulation sections, the fixed bit rate, and the moving picture encoding accumulated in the accumulation section Each parameter used by the quantization subsection is determined by the difference between the number of bits of the signal and the unoccupied capacity of the input buffer obtained by the virtual occupied capacity calculation section.

In addition, the moving image encoding method of the present invention is based on the sum of the bit numbers of the plurality of moving images generated by encoding the plurality of moving image signals and outputting each of the plurality of moving image encoding signals after temporary accumulation. The difference between the output bits simultaneously encodes a plurality of moving image signals using each preset quantization width. The method includes the steps of calculating occupancy of a plurality of encoding buffers of temporally accumulated moving image encoding signals, calculating when the moving image encoding signals are not input to the input buffer after a preset time period for the preset time period each of the unoccupied capacity of a plurality of input buffers of the decoder decoding device; A buffer difference is obtained that is less than the preset value using a quantization width smaller than the corresponding preset quantization width.

Therefore, even when the number of bits of the moving image signal suddenly increases due to a corresponding portion of an image being extremely complex, it is possible to avoid the occurrence of a problem in the input buffer when the moving image coded signal is decoded by the decoding means. Underflow occurs in .

In an embodiment of the present invention, the preset value is 0.

In addition, the moving image encoding method of the present invention is based on the sum of the bit numbers of the plurality of moving images generated by encoding the plurality of moving image signals and outputting each of the plurality of moving image encoding signals after temporary accumulation. The difference between the output bits simultaneously encodes a plurality of moving image signals using each preset quantization width. The method includes the steps of: calculating occupancy of a plurality of encoding buffers of temporally accumulated motion picture encoding signals, calculating after a preset time period when the motion picture encoding signals are not input to the input buffer during the preset time period each of the unoccupied capacity of a plurality of input buffers of the decoder decoding device; A buffer difference is obtained that is greater than a preset value using a quantization width that is larger than the corresponding preset quantization width.

In an embodiment of the present invention, the preset value is 0.

In addition, the moving image encoding method of the present invention is based on the sum of the bit numbers of the plurality of moving images generated by encoding the plurality of moving image signals and outputting each of the plurality of moving image encoding signals after temporary accumulation. The difference between the output bits simultaneously uses each preset quantization width to encode a plurality of moving image signals, comprising the steps of: calculating the occupancy of a plurality of encoding buffers of temporarily accumulated moving image encoding signals, calculating The unoccupied capacity of each of the plurality of input buffers of the decoding device after a predetermined time period and when the motion picture coded signal is not input to the input buffer for the predetermined time period; A buffer difference obtained by subtracting the unoccupied capacity of the input buffer from the occupation of the coding buffer of a moving image coding signal, and adding a difference of the buffer greater than a preset value to the sum of the bits sum of values.

In an embodiment of the present invention, the preset value is 0.

According to still another aspect of the present invention, there are provided a moving picture coding device and a multiplexing device for coding a plurality of moving picture signals and multiplexing the coded signals. The device includes: a moving image switch section for multiplexing a target of a plurality of input moving images; a plurality of coding sections for encoding a plurality of input moving images output from the moving image switching section; a multiplexing section that multiplexes coded signals output from a plurality of coding sections; A control section for multiplexing mode; wherein the control section switches the encoding mode of the encoding section and the multiplexing mode of the multiplexing section according to the switching state of the moving picture switching section.

With the above constitution, a plurality of moving picture signals having different precisions can be encoded, multiplexed and outputted simultaneously. At this time, the combination of a plurality of input moving picture signals of respective precision levels is not limited. This simultaneous encoding and multiplexing is also applicable in the case where all input moving picture signals are independent of each other or all input moving picture signals are generated by screen division. In addition, even when the output bit rate is fixed after the multiplexing, it is possible to simultaneously encode, multiplex and output a plurality of individual moving picture signals and Multiple motion picture signals.

In one embodiment of the present invention, the control section includes: a switch control subsection for switching the switching state of the moving image switch section; a coding control subsection for switching the coding mode of the coding section; and A multiplexing control subsection to switch the multiplexing mode of the multiplexing section, and the encoding control subsection switches the encoding mode according to the switching state of the moving picture switch section, the multiplexing The multiplexing mode is switched by the control subsection according to the state of the motion picture switching section.

In still another embodiment of the present invention, the moving picture encoding and multiplexing apparatus further includes a screen dividing section upstream of the moving picture switching section.

In still another embodiment of the present invention, the moving picture encoding and multiplexing apparatus further includes a moving picture time division section upstream of the moving picture switching section.

In yet another embodiment of the present invention, the motion picture coding and multiplexing apparatus further includes an accumulating section for temporarily accumulating the motion picture coded signal output from the coding section, wherein the coding control section according to the A parameter for controlling each encoding section is detected by the sum of the bits of the motion picture encoded signal accumulated by the accumulating section.

In yet another embodiment of the present invention, the multiplexing section includes: a motion picture multiplexing subsection for multiplexing the motion picture coded signal and outputting the motion picture coded signal; Multiplexing a plurality of input signals input to the multiplexing section and a path multiplexing subsection of output signals output from the moving image multiplexing subsection; A switching subsection of the object of the moving picture coded signal of the multiplexing section between the moving picture multiplexing subsection and the path multiplexing subsection.

According to still another aspect of the present invention, an image transmission device is provided. This apparatus includes a plurality of image encoding sections and a transmission processing section, wherein the transmission processing section multiplexes the encoded data specified by the image processing section at each multiplexing timing, and transmits the encoded data to a preset route, and send to the image encoding section the encoded data amount calculated based on the encoded data amount in the image encoding section and can be used to calculate a value until the encoded data is transmitted to a decoding device The amount of delay, and the plurality of image encoding sections determine the quantization width used in quantization in an image encoding process based on the multiplexing information sent from the transmission processing section, and in each multiplexing The coded data is transmitted to the transmission processing section at timing to specify the coded data.

In an embodiment of the present invention, the transmission processing section includes a transmission buffer, and the transmission processing section receives encoded data specified by the image encoding section in the transmission buffer every multiplexing timing, and outputting the encoded data at a preset transmission rate in the order in which the encoded data is received, and outputting an occupancy of the encoded data in the transmission buffer as multiplexed information.

In one embodiment of the present invention, each image encoding section includes: a quantization width determination subsection; and an image for encoding the image quantized by the quantization width specified by the quantization width determination subsection The basic encoding processing subsection of the quantization width determination subsection calculates, based on the multiplexing information, the number of each of the plurality of decoding devices that receive the transmitted encoded data from each image encoding section through the transmission processing section. a buffer occupancy, and transmit the quantization width based on the calculated buffer occupancy in the decoding device and the amount of coded data not transmitted.

In another embodiment of the present invention, each image coding section further includes a coded data storage subsection for temporarily storing the coded data, and the coded data storage subsection temporarily stores each transmitted Timing the non-transmitted data which is not transmitted to the transmission processing section among the encoded data, and using a value preset by each image encoding section as an upper limit or by The coded data is transferred to the transfer processing section from a value obtained by subtracting the occupancy of the buffer in the decoding device calculated based on the multiplexing information from a buffer size of the decoding device shown.

In addition, the image transmission apparatus of the present invention includes a transmission processing section for receiving encoded data from a plurality of image encoding sections, wherein the transmission processing section includes a transmission buffer, and at each multiplexing timing The coded data specified by the image coding section is received in the transfer buffer, and the coded data is output at a preset transfer rate in the order in which the coded data is received and output as multiplexed information occupancy of the encoded data in the transmit buffer.

In addition, the image encoding apparatus of the present invention includes a quantization width determining section and a basic encoding processing section for encoding an image of an image quantized using a quantization width specified by the quantization width determining section, wherein The quantization width determining section is based on an amount of delay available until encoded data is transferred to a virtual decoding device, occupancy of a buffer in the virtual decoding device calculated based on the multiplexing information , and the calculated multiplexed signal of the amount of coded data not transmitted to determine the quantization width.

In an embodiment of the present invention, the image encoding device further includes an encoded data storage section for temporarily storing the encoded data, wherein the encoded data storage section temporarily stores the encoded data at each transmission timing. The untransmitted data that is not transmitted in the data, and use as an upper limit a value preset by each picture encoding part or subtract from a buffer size of the decoding device indicated by the receiving end according to A value obtained in the virtual decoding means calculated by the multiplexing information is output to output the coded data.

In addition, the image transmission device of the present invention includes a bit rate management section and one or more image encoding sections for managing a bit rate on a transmission line; wherein the bit rate management section determines the The bit rate of the sum of the transfer rates in one or more image coding sections on the above, for the transmission of the required coded data to be transmitted by any image coding section at each transmission timing, and the transmission and Transmission information about a transmission delay as a function of another allowed image coding portion on the transmission route, and the one or more image coding portions are coded by using untransmitted codes according to the transmission information The data and the image of the image quantized by a quantization width determined by a given buffer size at the receiving end.

With the above constitution, the quality of image signals transmitted over a network can be substantially uniform.

In addition, the image transmission apparatus of the present invention includes a bit rate management section for managing the transmission rates of a plurality of image coding sections, wherein the bit rate management section determines the A bit rate of the sum of transmission rates for the transmission of the coded data required to be transmitted by any picture coding part at each transmission timing, and sent with another allowed picture on the transmission line Transmission information related to transmission delay as a part of encoding.

In addition, the image coding apparatus of the present invention sets a user on the basis of transmission information related to a transmission delay varying with an allowable transmission amount, untransmitted coded data, and a buffer size indicated at the receiving end. Quantization width for quantization.

In addition, the image encoding method of the present invention includes the steps of: (1) determining a quantization width; and (2) encoding an image using the image quantized by the quantization width specified in the step (1), wherein in the step The quantization width in (1) is set based on multiplexing information that can be used to calculate an amount of delay obtained until encoded data is transferred to a virtual decoding device in which A buffer occupancy is calculated based on the multiplexing information and an untransmitted coded data amount.

In one embodiment of the present invention, untransmitted data that is not transmitted in the coded data is temporarily stored at each transmission timing, and the coded data utilizes a value preset by each image coding section as an upper limit. A value or a value obtained by subtracting occupancy of the buffer in the decoding device calculated from the multiplexing information from the buffer size of the decoding device indicated by the receiving end is used output.

In addition, the image coding method of the present invention includes the step of setting a user by setting a transmission delay based on transmission information related to a transmission delay with a variation of an allowable transmission amount, untransmitted coded data, and a predicted buffer size at the receiving end. A picture is coded at a quantization width of quantization.

Therefore, the present invention has the advantages: (1) Provides a moving picture encoding device, method capable of outputting a moving picture with uniform quality and capable of preventing the picture even after a scene change when the picture suddenly becomes complicated. (2) provided moving image multiplexing device/method can multiplex moving image signals and output its result so that each input buffer of a decoding device does not appear Overflow or underflow; (3) provided moving image coding and multiplexing device/method can simultaneously encode moving image signals with different resolutions, multiplex the coded signals, and output As a result; and (4) the provided image transmission device can maintain uniformity in the quality of the encoded signal by the encoding section even when a scene changes, and, in an application where encoders are distributed in a network, The processor distributes the rate and obtains a uniform image quality according to the input image over the entire allowed bit rate range.

Description of drawings

These and other advantages of the present invention will be more clearly understood to those of ordinary skill in the art after reading and understanding the following detailed description in conjunction with the accompanying drawings.

Fig. 1 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 1 of the present invention.

FIG. 2 is a block diagram of an encoding portion of the apparatus of FIG. 1. FIG.

FIG. 3 is a block diagram of a multiplex control portion of the apparatus of FIG. 1. FIG.

Figure 4 shows the quantization width with respect to the number of bits in the accumulation section.

Fig. 5 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 2 of the present invention.

FIG. 6 is a block diagram of a multiplexing portion of the apparatus of FIG. 5. FIG.

FIG. 7 is a flowchart of the operation of the apparatus of FIG. 5. FIG.

FIG. 8 is a flowchart at step 204 in FIG. 7 .

FIG. 9 is a flowchart at step 208 in FIG. 7 .

Fig. 10 is a flow chart showing the operation of a moving picture encoding and multiplexing apparatus according to Example 3 of the present invention.

11A to 11C show the processing at step 502 in FIG. 10, where the lower limit of the transfer amount is calculated.

Fig. 12 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 4 of the present invention.

FIG. 13 is a block diagram of the encoding portion of FIG. 12 .

14A and 14B show calculations performed by the decoder buffer unoccupied capacity calculation section of the apparatus of FIG. 12. FIG.

FIG. 15 is a flowchart of the operation of a stuffing control portion of the apparatus of FIG. 12. FIG.

Fig. 16 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 5 of the present invention.

Fig. 17 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 7 of the present invention.

Fig. 18 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 8 of the present invention.

FIG. 19 shows a screen splitting method used by the apparatus of FIG. 18. FIG.

Fig. 20 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 9 of the present invention.

FIG. 21 is a block diagram of a multiplexing portion of the apparatus of FIG. 20. FIG.

Fig. 22 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 10 of the present invention.

Fig. 23 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 11 of the present invention.

Fig. 24 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 12 of the present invention.

Fig. 25 is a block diagram of an image transmission apparatus according to Example 13 of the present invention.

FIG. 26 is a block diagram of an encoding subsection of an image encoding section of the apparatus of FIG. 25. FIG.

FIG. 27 is a flowchart of the operation of the encoding subsection of FIG. 26 .

Fig. 28 is a block diagram of an image transmission apparatus according to Example 14 of the present invention.

Fig. 29 shows a header format of a packet used in the Fig. 28 device.

FIG. 30 shows a communication method between two nodes of the apparatus of FIG. 28 .

FIG. 31 shows the transmission of a packet used in the apparatus of FIG. 28 for notifying the number of packets.

Fig. 32 shows the transfer of a packet for image transfer used in the Fig. 28 device.

FIG. 33 shows the transmission of packets generated by the bit rate management node of the apparatus of FIG. 28. FIG.

FIG. 34 is a block diagram of a node of the apparatus of FIG. 28 .

35A and 35B are graphs comparing bit rates used between a conventional device and the device of FIG. 28. FIG.

Figure 36 is a diagram of an embodiment of the present invention using a general computer.

Fig. 37 is a block diagram of a conventional moving picture encoding device.

Fig. 38 is a block diagram of a conventional image transmission device.

Detailed ways

The invention will be illustrated by means of the following examples with reference to the accompanying drawings.

(example 1)

Fig. 1 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 1 of the present invention.

The motion picture encoding and multiplexing device of this example comprises: compression/encoding is transformed by signal and quantized coding part 1-1A to 1-NA of the input motion picture (wherein N is an integer); Accumulation sections 2-1 to 2-N of moving picture coded signals output from parts 1-1A to 1-NA; multiplex the moving picture coded signals at a fixed bit rate and output the multiplexed result Multiplex section 3 of signal; Control the quantization control section 4 of a quantization width used when being quantized by this coding section 1-1A to 1-NA; The decoding delay time calculation section 5 of the time period (decoding delay time) until this bit is decoded by the decoding device; The frame number calculation section 6 of the number of bits in each frame that the coded signal is composed of; the multiplexing control section 7A that controls the multiplexing section 3; An adder 8 that adds the number of bits of the motion picture coded signal accumulated in it.

FIG. 2 is an enlarged view of the encoding section 1-1A shown in FIG. 1. Referring to FIG. When a moving picture signal representing a moving picture is input to the encoding section 1-1A, the differentiator 101-1 calculates the input moving picture and the accumulated moving picture of one frame of a frame memory 108-1 The difference between them, and output a difference motion picture. The differential motion picture is then transformed into -DCT (Discrete Cosine Transform) coefficients by a DCT subsection 102-1. The DCT coefficients are then quantized by a quantization subsection 103-1 with a quantization width specified by the quantization control section 4 and converted into a motion picture coded signal by a variable length coding subsection 104-1. The resulting motion picture encoded signal is output from the encoding section 1-1A. The output signal from the quantization subsection 103-1 is also sent to the inverse quantization subsection 105-1, where the input signal is dequantized. The inverse quantized signal is transformed into a decoded differential motion picture by the inverse DCT subsection 106-1. The decoded differential motion signal is added to the motion image of one frame accumulated in the frame memory 108-1, and is converted into a motion image decoded by the decoding means. The resulting moving pictures are accumulated in the frame memory 108-1.

The decoding delay time calculation section 5 calculates the delay of the bit of the motion coded signal accumulated in each of the accumulation sections 2 - 1 to 2 -N which is initially set to the multiplexing section 3 . The decoding delay time of the bit initially output from each of the encoding sections 1-1A to 1-NA is used as an initial value. This is preset according to the size of an input buffer of the decoding device, the output bit rate at the multiplexing section, and the like. After that, the value is updated according to the decoding process in the decoding means.

The frame bit number calculating section 6 calculates the bit number of one frame including the bits of the motion coded signal accumulated in each of the adding sections 2 - 1 to 2 -N initially set to the multiplexing section 3 .

The multiplexing control section 7A repeats the calculation of the number of bits of the moving image coded signal output from each of the accumulation sections 2-1 to 2-N to the multiplexing section 3 during a predetermined period.

The multiplexing section 3 multiplexes the moving image encoded signal according to the number of bits calculated by the multiplexing control section 7A, and outputs the multiplexed signal at a fixed bit rate.

FIG. 3 is an enlarged view showing the configuration of the multiplexing control section 7A in FIG. 1. Referring to FIG. The first parameter generation subsection 71 receives a frame of bits including an initial setting to the multiplex with a decoding delay time for each of the accumulation sections 2-1 to 2-N. The accumulated bits of section 3 are multiplexed, and the number of bits of the frame is divided by the decoding delay time. The second parameter generation subsection 72 adds the values obtained by the first parameter generation subsection 71 for the accumulation sections 2-1 to 2-N. The third parameter subsection 73 divides the values obtained by the second parameter generation subsection 72 by the respective values obtained by the first parameter generation subsection 71 for the accumulation sections 2-1 to 2-N. When the number of bits output from the accumulation sections 2-1 to 2-N to the multiplexing section 3 has been calculated, the multiplexing number of bits calculation subsection 74 receives The bit number of the moving image coded signal of output multiplies the bit number received by the value obtained by the third parameter generation subsection 73, and notifies its result value of the multiplexing section 3 as the value from the accumulating section 2-1. Each of to 2-N is output to the number of bits of the multiplexing section 3.

The operation of the apparatus of Example 1 above was as follows. The moving image encoded signals output from the encoding sections 1-1A to 1-NA are temporarily accumulated in the accumulation sections 2-1 to 2-N. The number of bits of the moving image encoded signal accumulated in the accumulation sections 2-1 to 2-N is added by an adder 8, and the resulting value is sent to a quantization control section 4. The quantization control section 4 calculates a quantization width based on the received value, ie, the sum of the number of bits of the moving image coded signal, and the moving image is quantized based on the quantization width. The multiplexing control section 7A calculates the number of bits output from the accumulation sections 2-1 to 2-N to the multiplexing section 3 using the values obtained by the decoding delay time calculation section 5 and the frame bit number calculation section 6 . The multiplexing section 3 multiplexes the moving image encoded signal using the number of bits obtained by the multiplexing control section 7A, and outputs the resulting moving image encoded signal at a fixed bit rate.

In the above operation, the quantization control section 4 converts the sum of the bits of the moving image coded signal into a quantization width. FIG. 4 is a diagram showing an example of such conversion. The X-axis of this figure represents the sum of the bits of the moving image coded signals accumulated in the accumulation sections 2-1 to 2-N, which is equal to the sum of the bits of the moving image coded signals produced by encoding The difference between the number of bits of the moving image coded signals output from the accumulating sections 2-1 to 2-N of the bit rate is calculated. The Y-axis of the graph represents the quantization width. In this example, the quantization width is set larger as the number of bits is larger, and is set smaller as the number of bits is smaller.

The quantization control section 4 calculates the quantization width based on the characteristics shown in the graph of FIG. 4, and supplies the quantization width to all the coding sections 1-1A to 1-NA. In other words, all encoding sections 1-1A to 1-NA have the same value of quantization width and carry out the quantization according to this width.

As in Example 1 above, since a plurality of input moving pictures are compressed/encoded using the same quantization width, the quality of a plurality of output moving pictures can be made uniform. In addition, since the quantization control is performed based on the sum of the bits of the moving image coded signals accumulated in the accumulation sections 2-1 to 2-N, the bit rate can be effectively used.

Each moving picture input to the encoding sections 1-1A to 1-NA may constitute an independent moving picture or a group of parts obtained by separating a moving picture. In the latter case, each group is considered as a frame unit.

In Example 1, all coding sections 1-1A to 1-NA are controlled by the same quantization width. In addition, different quantization widths can be utilized by providing a limitation such as setting the priority and An upper limit is set on the output bit rate.

(Example 2)

Fig. 5 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 2 of the present invention.

In Fig. 2, there are two parts to be additionally provided to the device of example 1 shown in Fig. 1: that is, a decoder buffer that determines the unoccupied capacity of the decoder buffer of the input buffer of the decoding device. An occupied capacity calculation section 9, and an accumulation bit calculation section 12 which calculates the respective number of bits accumulated in the accumulation sections 2-1 to 2-N.

The decoder buffer unoccupied capacity calculation section 9 includes a timer, and receives bits output from each of the accumulation sections 2-1 to 2-N to the multiplexing section 3 from the multiplexing control section 7B. The received number of bits is added to the occupancy in the input buffer of a virtual decoding device for each accumulation section 2-1 to 2-N. This decoder buffer unoccupied capacity calculation section 9 also receives the number of bits input from the frame bit calculation section 6 into the input buffer of the decoder device in each frame and receives the information from the decoding delay time calculation Decoding delay time per frame of part 5. The time at which decoding begins for a frame is determined based on the decoding delay time for that frame. When the timer indicated that the decoding time was increasing, the decoder buffer unoccupied capacity calculation section 9 subtracted the number of bits of the frame decoded from the occupied amount in the input buffer of the decoding device , subtract the occupancy of the result obtained in the input buffer from the capacity of the input buffer, and output the result value to the multiplexing control section 7B as the current unoccupied input buffer of the decoding device capacity.

FIG. 6 shows an exemplary configuration of the multiplexing control section 7B. A minimum value selection subsection 75 is additionally provided downstream of the multiplex bit number calculation subsection 74 of the multiplex control section 7A of Example 1 shown in FIG. The minimum value selection subsection 75 selects the current decoder buffer unoccupied capacity obtained from the product obtained by the multiplexing bit calculation section 74, the decoder buffer unoccupied capacity calculation section 9, and The minimum value among the accumulation number of bits in each accumulation section obtained by the accumulation number of bits calculation section 12, and the resultant value is sent to the multiplexing section 2 as output from the accumulation section to the multiplexing section 3 (hereinafter, such a number of bits is referred to as a multiplexed number of bits).

FIG. 7 is a flow chart of the operation of the multiplexing control section 7B for determining the number of bits of the moving picture coded signal output from each of the accumulation sections 2-1 to 2-N to the multiplexing section 3. .

First, the multiplexing control section 7B calculates the output bit rate from the multiplexing section 3 during each calculation of the multiplexing control section 7B based on a preset output bit rate in the multiplexing section 3. The number of bits of the output video coded signal (step 201). Then, the decoding delay time for each of the accumulation sections 2-1 to 2-N is received from the decoding delay time calculation section 5 (step 202). The number of bits in each frame for each of the accumulation sections 2-1 to 2-N is obtained from the frame bit number calculation section 6 (step 203). A best multiplexed number of bits for each accumulation section 2-1 to 2-N is calculated in the procedure described later (step 204). Then, the number of bits of the moving image coded signal accumulated in each of the accumulation sections 2-1 to 2-N is obtained from the accumulation bit calculation section 12 (step 205). Afterwards, for each accumulation section 2-1 to 2-N by subtracting the capacity of the output buffer in the input buffer of the decoding device from the capacity of the output buffer obtained by the decoder buffer unoccupied capacity calculation section 9 The unoccupied capacity of the input buffer is calculated by accumulating the number of bits (step 206). Next, select as the second multiplexing number among the initial multiplexing bit number obtained in step 204, the number of bits obtained in step 205, and the unoccupied capacity in the input buffer obtained in step 206 The minimum value of multiplexing bits (step 207). Finally, the number of bits of the moving image coded signal output from each accumulation section 2-1 to 2-N to the multiplexing section 3 is calculated in a step described later (step 208). The above processing is repeated.

The flowchart of FIG. 8 shows in detail the procedure of step 204 shown in FIG. 7 for calculating the number of initial multiplexed bits.

First, the decoding delay time is divided by the number of bits of the motion coded signal of one frame accumulated in each accumulation section 2-1 to 2-N, and the resulting value is expressed by R _i (i=1 to N) To label (step 301). Then, the sum of the values R ₁ to _RN is calculated (step 302 ). Calculate the main multiplexing number of bits P1 _i for each motion coded signal by the following formula (1) (step (303)

Afterwards, among the initial multiplexed number of bits P1 _i , the number of bits obtained in step 205, and the unoccupied capacity of the input buffer obtained in step 206, the one for each accumulation section 2-1 to 2-N is selected. The minimum value is used as the second multiplexing bit number P2 _i .

9 is a flow chart showing in detail the processing of step 208 in FIG. 7 for calculating the number of bits output from each accumulation section 2-1 to 2-N to the multiplexing section 3.

First, if the main multiplex number of bits P1 _i and the sub-multiplex number of bits P2 _i are the same for all moving picture coded signals accumulated in the accumulation sections 2-1 to 2-N , the processing is terminated (step 401). Otherwise, it is determined whether the number of bits P1 ₁ and P2 _i are the same for each motion picture coded signal (step 402). If they are the same, in addition to the sub-multiplexed number of bits P2 _i obtained in step 207, a multiplexed number of bits P3 _i is calculated using the following equation (2) (step 403).

P3 _i = minimum (encoder buffer occupation, encoder buffer to occupy capacity) -

P2 _i (2)

Here, "minimum" means a smaller value used between the number of bits obtained in step 205 and the unoccupied capacity of the input buffer obtained in step 206 .

If the number of bits P1 _i and P2 _i are different, set the number of bits P3 _i to 0 (step 404). Then, a short number of bits D obtained in step 207 for the sum of the second multiplexed bits P2 _i of the accumulation parts 2-1 to 2-N is calculated by the following formula (3) (Comparison of the number of bits output from the multiplexing section 3 at step 405).

D＝Output digits-∑P2 _i (3)

After that, the sum S of the number of bits P3 _i obtained in steps 403 and 404 is calculated (step 406). Determine whether the sum S is greater than the short number of digits D (step 407). If yes, place as the number of multiplexed bits of the motion coded signal obtained by the following equation (4) (step 408). If not, it is set to the multiplexed number of bits of the motion coded signal obtained by the following formula (5) (step 409).

Number of multiplexing bits = P2 _i + P3 _i

Therefore, in Example 2, the moving picture codes accumulated in the accumulation sections 2-1 to 2-N are controlled according to the number of bits in each frame and the time at which each frame is input to the corresponding decoding means. Multiplexing of signals. Therefore, underflow in the input buffer of the decoding device is prevented. In addition, the number of bits accumulated in the input buffer of each decoding means is counted to control multiplexing so that bits exceeding the unoccupied capacity of the input buffer are not transmitted. Thus, overflow in the input buffer is prevented.

(Example 3)

Fig. 10 shows a flowchart illustrating the operation of a moving picture encoding and multiplexing apparatus according to Example 3 of the present invention. The configuration of the apparatus of Example 3 is the same as that of the apparatus of Example 1 shown in FIG. 1, but the processing of the apparatus of Example 1 is different.

The flowchart of FIG. 10 shows the motion diagram of the multiplexing control section 7A determining the multiplexing bit number, that is, output to the multiplexing section 3 from each accumulation section 2-1 to 2-N. Like the operation of encoding the number of bits in a signal.

First, the multiplexing control section 7A calculates the motion output from the multiplexing section 3 during each calculation period of the multiplexing control section 7A based on a preset output bit rate at the multiplexing section 3. The total number of bits of the image coded signal (step 501). Then, the lower limit of the transmission amount of the moving picture coded signal for each accumulation section 2-1 to 2-N in the step described later is calculated, and the lower limit bit of the moving picture The coded signal is output to the multiplexing section 3 from each of the accumulation sections 2 - 1 to 2 -N. In order to compare the frames sequentially used in each accumulation section 2-1 to 2-N in such a manner that the frame with the smaller decoding delay time is selected earlier, an initial value for a reference time is set (step 503). The number of bits in a head frame in each of the accumulation sections 2-1 to 2-N is read from the frame bit number calculation section 6 (step 504). The decoding delay time of the head frame in each accumulation section 2-1 to 2-N is read from the decoding delay time calculation section 5 (step 505). The received decoding delay time of each frame is compared with the reference time (step 506). For the accumulation part of the accumulated frames with decoding delay time shorter than the reference time, first calculate the unoccupied capacity of the input buffer of the corresponding decoding device (step 507). Then, select among the maximum number of transmission bits determined by the number of bits of the frame obtained in step 504, the unoccupied capacity of the input buffer obtained by step 507, and the maximum output bit rate of the moving image coded signal The minimum value of is used as the initial number of multiplexed bits. For the accumulated portion of the accumulated frame having a decoding delay time greater than the reference time, the main multiplex bit is set to 0 (step 508). Then, the sum of the number of main multiplexing bits obtained for each of the accumulation sections 2-1 to 2-N is calculated. If the sum is 0, the period for calculating the number of multiplexed bits is determined (step 509). If the sum is less than the number of bits remaining after subtracting the lower limit obtained by step 502 from the total number of bits obtained by step 501, then the initial multiplexed bits obtained in step 508 The number is added to the number of multiplexing bits for each accumulation section 2-1 to 2-N (step 511). And, the reference time is updated, for example by adding the frame period time to the reference time (step 512). At this time, the moving picture coded signal of the main multiplexed number of bits is output to the multiplexing section 3 from each of the accumulation sections 2-1 to 2-N. If the sum of the number of multiplexing bits is greater than the reserved number of bits, the value obtained by proportionally allocating according to the main multiplexing number of bits is used as the secondary multiplexing number of bits (step 513), And the sub-multiplexed bits are added to the multiplexed bit (step 514), thereby terminating the cycle of the multiplexed bit calculation. At this time, the moving picture coded signal of the sub-multiplexed number of bits is output to the multiplexing section 3 from each of the accumulation sections 2-1 to 2-N. The above-mentioned processes are sequentially repeated.

The lower limit used to calculate the transfer amount shown in FIG. 10 will be described with reference to FIGS. 11A to 11C. FIG. 11C is an enlarged view of FIG. 11B .

The graphs in FIGS. 11A to 11C are plotted according to the number of bits per frame accumulated in the buffer (the accumulation section is located downstream of the buffer of the encoding section or decoding means) at a certain time. The x-axis in these figures is the time axis, representing the time period required for decoding one frame in the buffer. Each flat portion of the parameter X represents a frame period. The Y-axis represents the number of bits: a value below 0 indicates the number of bits of the moving image coding signal accumulated in the buffer of the decoding device, and a value equal to or greater than 0 indicates the number of bits accumulated in the buffer of the encoding device The number of bits of the motion picture coding signal.

Now, the broken line of the parameter X in Figs. 11A and 11B will be described. First, the time period required until an earliest decoded frame is decoded is determined and plotted at the X coordinate from origin 0, and the total number of bits accumulated in the buffer of the decoding device is determined and Draw with a Y coordinate below the origin 0, so you get point A. Using point A as a starting point, draw a line up the Y-axis with a length corresponding to the number of bits of the frame. Then, a straight line corresponding to the length of one frame period is drawn along the X-axis. Thereafter, a straight line having a length corresponding to the number of bits of a frame next to the earliest decoded frame is drawn along the X-axis, followed by a line corresponding to the period length of a frame. The rendering of frames in the buffer is repeated in such a way that earlier decoded frames are rendered earlier. For the currently encoded frame, draw a line according to the number of bits encoded. The endpoint of this polyline is marked by point B.

The polyline of the parameter Y is now described. Draw the line starting from point B indicating that the number of bits produced is the number of bits of the currently encoded frame, so that it is always above the parameter X and above the origin 0, and the maximum elevation angle does not exceed that used for motion picture coding and the maximum bit rate of the multiplexing device. The parameter Y represents the minimum number of bits required to be transmitted to the decoding device within a predetermined time period and avoid overflow at the decoding device. In this example, the time on the x-axis is assumed to be the elapsed time from the current time. The minimum number of transfer bits corresponding to the lower limit of the transfer amount is obtained by step 502 in each calculation cycle.

The lower limit of the minimum transfer amount in the next calculation cycle is 0 for the parameter Y in FIG. 11A, while it is L (L>0) for the parameter Y in FIG. 11B.

As described above, in Example 3, one frame whose decoding delay time is shorter is output earlier, and the number of bits necessary for avoiding overflow is preferentially output. Therefore, the overflow of the number of bits in the input buffer of the decoding device can be prevented.

(Example 4)

The arrangement of Example 1 above is effective when the minimum output bit rate and the capacity of the input buffer of the decoding device are not restricted. However, when they are limited, an overflow may occur in the input buffer of the decoding device because of the transmission delay of the motion picture coded signal. For example, assume that a sudden increase in the number of bits occurs in an encoding section because an input image section is very complex. In this case, in the decoding device that increases the number of decoding bits, the number of bits of the frame input to the input buffer for decoding becomes larger than the number of bits of the frame read from the input buffer, and as a result The input buffer is filled. Therefore, the number of bits in the accumulation section downstream of the encoding section corresponding to the sudden increase in the number of bits is not reduced. To compensate, the number of bits output from the accumulation sections located downstream of other encoding sections whose number of bits has not been increased is increased, with the result that the number of bits in these accumulation sections is reduced. As a result, the effective number of bits in all accumulation sections decreases, resulting in a shorter output bit rate in the multiplexing section 3 . Since the quantization control is carried out on the premise that the output bit rate of the multiplexing section 3 is always fixed, an overflow occurs in the input buffer of the decoding device due to a decrease in the amount of transfer. To overcome this problem, it is pre-required to stuff the signal with bits that do not affect the decoding.

Fig. 12 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 4 of the present invention. The device of Example 4 provides the following addition to the part of the Example 1 device shown in Figure 1: that is, for calculating the number of bits accumulated in the input buffer of the decoding device to determine the unoccupied A decoder buffer of capacity unoccupied capacity calculation section 9; One-digit difference calculation section 10, for each accumulation section 2-1 to 2-N, calculates the number of bits in the accumulation section and correspondingly lies in the result of the translation The difference between the unoccupied capacity of the buffer of the decoding device of the accumulation section obtained by the coder buffer unoccupied capacity calculation section 9; The difference controls a number of stuffing bits generated by each encoding section 1-1B to 1-NB.

FIG. 13 is an example configuration of encoding sections 1-1B to 1-NB in FIG. 12 . This configuration differs from that of Fig. 2 in that a stuffing sub-section 109-1 is provided downstream of the variable-length coding sub-section 104-1.

The stuffing subsection 109-1 inserts stuffing bits of a stuffing amount specified by the stuffing control section 11 into the signal output by the variable-length coding subsection, and outputs the resulting signal.

In the apparatus of this example having the above-mentioned constitution, the decoding delay time calculation section 5 calculates the motion map initially sent to the multiplexing section 3 for storage in each of the accumulation sections 2-1 to 2-N The decoding delay time of the bits of the coded signal is as described in the example, and the result is output to the decoder buffer unoccupied capacity calculation section 9. The frame number calculation section 6 calculates the number of bits of a frame including the bits of the moving image coded signal accumulated in each of the accumulation sections 2-1 to 2-N initially supplied to the multiplexing section 3, As described in the example, and the result is output to the buffer unoccupied capacity calculation section 9 in the decoder. The adder 8 adds the number of bits of the moving image coded signals accumulated in the accumulation sections 2-1 to 2-N, and outputs the result to the decoder buffer unoccupied capacity calculation section 9.

The decoder buffer unoccupied capacity calculating section 9 is provided with a timer, and reserves a moving picture output bit rate as a part of the output bit rate in the multiplexing section 3 which distributes the moving picture coded signal.

In more detail, the decoder buffer unoccupied capacity calculation section 9 is based on the decoding delay time for each accumulation section 2-1 to 2-N and the number of bits in one frame in the accumulation section and by the Calculate the current unoccupied capacity of the input buffer of the decoding device by counting the value specified by the time. Thereafter, according to the sum of the bits of the moving image signals accumulated in each accumulation section 2-1 to 2-N and the output bit rate for the moving image encoding signal output from the multiplexing section 3, The time T in all the moving picture coded signals accumulated in each of the accumulation sections 2-1 to 2-N is transmitted at the output bit rate calculated by the following equation (6).

Thereafter, the unoccupied capacity of the input buffer of the decoding device after time T has elapsed is calculated on the assumption that no motion picture coded signal is input to the input buffer of the decoder from the current time until time T.

The calculation by the decoder buffer unoccupied calculation section 9 will be described below with reference to Figs. 14A and 14B. Fig. 14A has shown the present state of the input buffer of accumulation part and decoding device, and wherein A1 shows the current number of bits in this accumulation part and B1 represents the unoccupied capacity calculation part 9 by the decoder buffer. The currently unoccupied capacity of the input buffer of the decoding device. Assume that the frames decoded by the decoding means during the period from the current time up to time T are three frames. Then, as shown in FIG. 14B, the decoder buffer unoccupied capacity calculation section 9 outputs a value B2 corresponding to the sum of the value B1 and the number of bits of the previous three frames in the input buffer.

The bit difference calculation section 10 calculates, for each of the accumulation sections 2-1 to 2-N, the number of bits in the accumulation section and the decoding corresponding to the accumulation section obtained by the decoder buffer unoccupied capacity calculation section 9. The device inputs the difference between the unoccupied capacities of the buffers and outputs the resulting N difference to the stuffing control section 11.

FIG. 15 shows the flow of the stuffing control section 11. As shown in FIG.

First, the N difference obtained by the bit difference calculation section 10 is denoted by G _i (i=1 to N), and S _p and S _m are set to 0 for calculating the difference i and S m (step 601) . Then, it is judged whether each of the differences G ₁ to G _i is 0 (step 602 ). If the difference is equal to or greater than 0, the difference is added to _Sp (step 603). If the difference is less than 0, then the absolute value is added to _Sm (step 604). Then, it is judged whether S _p is greater than 0. If S _p is greater than 0, then processing proceeds to the calculation of the number of stuffing bits in steps 605-611. If _Sp is 0, processing aborts without stuffing (step 612). Then, it is judged whether S _m is greater than S _p (step 605). If S _m is equal to or greater than S _p , it is judged whether each difference G ₁ to G _N is less than 0 (step 606 ). If yes, the number of stuffing bits used for the motion picture coding signal corresponding to the difference G _i is calculated by the following equation (7) (step 607). If not, then set the stuffing bit to 0 (step 608).

If S _m is less than S _p , it is judged whether each of the differences G ₁ to G _N is less than 0 (step 809 ). If so, the number of stuffing bits used for the motion picture coded signal corresponding to the difference value G _i is set as the absolute value of G _i (step 610). If not, then the stuffing bit is set to 0 (step 611).

After obtaining the number of stuffing bits as described above, the stuffing control section 11 supplies the number of stuffing bits to the stuffing subsection 109-1 of the encoding section. The stuffing subsection 109-1 then stuffs the output signal with the number of stuffing bits provided.

Therefore, in the apparatus of Example 4 having the stuffing characteristic, when the number of bits of the portion of the moving picture coded signal generated is suddenly increased, overflow that may occur when stuffing is not performed is prevented.

The calculation method of the number of stuffing bits described above is just an example. Other methods can also be used. For example, the number of stuffing bits etc. can be increased or decreased for bytes aligned with the moving picture coded signal. The total number of stuffing bits can be increased or decreased by multiplying _Sp by a series. Alternatively, S _p may be equal to a moving image coded signal allocated to which S _m is less than 0.

In this example, whether or not the stuffing is performed is determined by whether or not the value obtained by the bit difference calculation section 10 is equal to or greater than 0. Any value other than 0 can be used.

(Example 5)

Fig. 16 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 5 of the present invention.

The apparatus of this example is the same as that of Example 4 shown in FIG.

The operation of the apparatus of this example having the above-mentioned constitution will now be described. When the conditions for stuffing are satisfied, the quantization control section 4 outputs a reduced quantization width to the encoding section on which stuffing has been performed among the encoding sections 1-1B to 1-NB. In more detail, when the quantization width measured by the quantization control section 4 based on the number of bits output from the adder 8 is Q, a quantization width [K×Q] is provided to the coding section in which stuffing has been performed, where K is a positive number less than 1 and [K×Q] is the largest integer within the range not exceeding K×Q. The quantization control section 4 receives the buffer difference value from the bit difference calculation section 10 and performs an operation similar to that shown in FIG. 15 . However, in this case, the quantization width is reduced in steps 607 and 610, while the quantization width is kept unchanged in steps 608 and 611.

Therefore, overflow in the decoding apparatus can be prevented by reducing the quantization width used for encoding in Example 5 to increase the number of bits instead of stuffing in Example 4.

(Example 6)

A moving picture encoding and multiplexing apparatus according to Example 6 of the present invention will now be described. The apparatus of Example 6 is the same as that of Example 5 except that the quantization control section 4 provides an increased quantization width to the encoding sections that do not satisfy the stuffing condition among the encoding sections 1-1B to 1-NB. In more detail, when the quantization width calculated by the quantization control section 4 is Q based on the number of bits output from the adder 8, a quantization width [K*Q] is supplied to the coding section where the stuffing is not performed, where K is a positive number greater than 1 and [K×Q] is the largest integer within the range of K×Q.

The quantization control section 4 receives the buffer difference value from the bit difference calculation section 10 and performs an operation similar to that shown in FIG. 15 . In this case, however, the quantization width is increased in steps 607 and 610, while it remains unchanged in steps 608 and 611.

Therefore, in Example 6, contrary to the case of Example 5, the quantization width is increased to reduce the number of bits for the coding portion in which stuffing is not performed, thus obtaining the same effect as described in Example 4. The operations of Examples 5 and 6 can be combined so that the quantization width of the encoded part whose median is increased is increased while the quantization width of the other encoded parts is decreased.

(Example 7)

Fig. 17 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 7 of the present invention. Except that the stuffing control section 11 is not provided and the multiplexing control section 7A performs control of the required number of stuffing bits for the multiplexing section 3 and outputs the stuffing control to the quantization control section 4 The apparatus of FIG. 17 is the same as that of FIG. 16 except that part 11 indicates the number of stuffing bits inserted by the multiplexing section 3.

In the apparatus of this example having the above-mentioned constitution, based on the number of bits in the accumulation sections 2-1 to 2-N commonly added by the adder 8 and the calculation result by the number of bits difference calculation section 10, the quantization control section 4 Computes the quantization width. In this example, the quantization control section 4 executes the processing in steps 601 to 604 for calculating _Sp in FIG. 15, and adds the obtained result _Sp to the number of bits in the accumulation sections 2-1 to 2-N Sum. In other words, the sum of the controlled stuffing bits inserted in the coding section in Example 4 is added to the sum of the bits of the actually generated motion picture coded signal. Then, the quantization width is determined as in Example 1. The required stuffing is carried out by the multiplexing section 3 under the control of the multiplexing control section 7A. When the sum of the total number of multiplexed bits is shorter than the sum of the number of bits accumulated in the accumulation parts 2-1 to 2-N due to preventing overflow in the buffer of the decoding device or the number of bits accumulated in the accumulation parts 2-1 to 2-N This stuffing results in compensation for a short number of bits when the number of bits output by section 3 is multiplexed.

In this example, the multiplexing control section 7A performs stuffing of the required number of bits. The quantization control section 4 determines the number of stuffing bits and adjusts the quantization size so that underflow in the decoding device is avoided. An error is generated between the judged number of stuffing bits and the actual number of bits, and the judged number of bits is adjusted accordingly. In order to adjust the judged number of bits, the quantization control section 4 receives the number of stuffing bits from the multiplexing control section 7A, and compares it with _Sp . Then, S _p is adjusted so that the error between the accumulated value of S _p and the accumulated value of the number of stuffing bits is within a predetermined range.

Therefore, in the apparatus of Example 7, the same effect can be obtained by a simpler constitution than that of the apparatus of Example 4.

(Example 8)

When the number of pixels constituting an input moving image is large, the processing rate of an encoding device may not be appropriate. Mask division coding is a method for coding such a moving picture signal of a moving picture composed of a large number of pixels.

In Examples 1 to 6, each encoding section mainly encodes a single moving image. The invention is also applicable to screen split coding.

Fig. 18 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 8 of the present invention. The apparatus of this example is provided with a screen dividing section 12 upstream of the encoding sections 1-1A to 1-NA of the apparatus of Example 1 shown in FIG.

The screen dividing section 12 divides an input moving picture into a plurality of segment screens as shown in FIG. 19, and assigns moving picture signals representing the respective segment screens to the encoding sections 1-1A to 1-NA. The encoding sections 1-1A to 1-NA encode the received moving image signal according to the quantization width supplied from the quantization control section 4. The same quantization width is provided to the encoding sections 1-1A to 1-NA. The resulting motion picture encoded signals output from the encoding sections 1-1A to 1-NA are temporarily accumulated in the accumulation sections 2-1 to 2-N. The number of bits in each of the accumulation sections 2 - 1 to 2 -N is collectively added by the adder 8 , and the added number of bits is output to the quantization control section 4 . The quantization control section 4 calculates a quantization width from the number of bits received from the adder 8 .

The frame bit number calculation section 6 calculates the number of bits of a frame including the bits of the moving image coded signal accumulated in each of the accumulation sections 2-1 to 2-N which are initially sent to the multiplexing section 3. The multiplexing control section 7A controls the multiplexing section 3 so that the moving image coded signal of the number of bits calculated by the frame bit number calculation section 6 is transferred from each accumulation section 2 in an order corresponding to the adopted screen division method. -1 to 2-N are sent to the multiplexing section 3. The multiplexing section 3 multiplexes the moving picture coded signal according to the control by the multiplexing control section 7A so that the coded signal constitutes a screen, and outputs the multiplexed signal at a fixed bit rate.

Therefore, in this example, moving pictures are encoded with the same quantization width for all segment screens at the time of this screen division encoding. This suppresses variations in image quality within a screen.

In this example, a screen is divided into four horizontally. In addition, the screen can be divided into several segments vertically or in a combination of horizontal and vertical directions. Also, the number of segments is not limited to four, and may be any natural number.

(Example 9)

Fig. 20 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 9 of the present invention.

Referring to Fig. 20, this device comprises: a motion picture switch part 1001, is used for selecting a motion picture signal from a plurality of input motion picture signals and switches the target of the selected motion picture signal; Coding part 1002- 1 to 1002-N (N is an integer greater than or equal to 1), which is used to encode an input moving image signal; a multiplexing section 1003; and a control section 1004, which is used to switch the switch of the moving image switching section 1001 The state, the encoding method of the encoding sections 1002-1 to 1002-N, and the multiplexing method of the multiplexing section 1003.

In the apparatus constructed as described above, the moving picture switching section 1001 receives moving picture signals or partial moving picture signals representing a plurality of segment screens obtained by dividing a screen of a high-precision moving picture. The control section 1004 determines whether each moving image signal input to the moving image switching section 1001 is output, and if outputted, determines which coding section the moving image signal will be output, and controls the moving image signal according to this judgment. Like switch part 1001. When a partial moving picture signal is input, all partial moving picture signals from the same moving picture signal are selected and output. The control section 1004 determines which of the moving image signal or partial moving image signal is to be encoded for each of the encoding sections 1002-1 to 1002-N in accordance with switching controlled by the moving image switching section 1001. The encoding sections 1002-1 to 1002-N encode the input moving image signal or part of the moving image signal and output the resulting encoded signal to the multiplexing section 1003. The control section 1004 controls the multiplexing section 1003 according to switching controlled by the moving picture switching section 1001 to multiplex the input moving picture signal. In the case of the screen division encoding, the multiplexing section 1003 multiplexes a part of the moving picture signal from the same moving picture signal, and then multiplexes the resultant part together with other multiple moving picture signals. The signal is multiplexed, and the resulting signal is output.

FIG. 21 is a block diagram of the multiplexing section 1003 of the apparatus of this example shown in FIG. 20 .

Referring to Fig. 21, this multiplexing part 1003 comprises: a moving image multiplexing subsection 1021, which is used for multiplexing the partial moving image coding signal obtained by encoding the moving image signal; The channel multiplexing subsection 1022 is used to multiplex a plurality of moving picture encoded signals; and switches 1023-1 to 1023-N (N is an integer greater than or equal to 1) are used to put The object of the coded signal is input to this multiplexing section 1003 .

The control section 1004 switches over the switches 1023-1 to 1023-N according to switching controlled by the moving picture switching section 1001. That is, the switch 1023 receiving part of the moving picture signal is switched to the moving picture multiplexing subsection 1021, while the other switches 1023 are switched to the channel multiplexing subsection 1022. The moving picture multiplexing subsection 1021 multiplexes a partial moving picture encoded signal obtained by encoding a partial moving picture signal from the same moving picture signal, and outputs the result. The channel multiplexing subsection 1022 multiplexes the moving picture coded signal output from the moving picture multiplexing subsection 1021 and from the switch 1023, and outputs the result thereof.

Therefore, in this example, a plurality of different moving picture signals and a plurality of partial moving picture signals generated by dividing a high-precision moving picture signal are encoded and multiplexed at once. In the apparatus of this example, coding and multiplexing can be performed even when all input signals are individual moving picture signals or they are partial moving picture signals generated by screen division.

In this example, a plurality of moving picture signals are separately divided and encoded, and the moving picture multiplexing subsection 1021 multiplexes a partial moving picture corresponding to one moving picture signal of all moving picture signals. like coded signals. In addition, a plurality of moving picture multiplexing subsections may be provided so that each moving picture multiplexing subsection multiplexes a set of partial moving picture coded signals.

(Example 10)

Fig. 22 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 10 of the present invention. In the apparatus of FIG. 10, the control section 1004 in Example 9 shown in FIG. 20 is divided into a switch control section 1041, an encoding control section 1042, and a multiplexing control section 1043.

The switch control section 1041 determines whether each moving image signal input to the moving image switching section 1001 is output, and if output, determines which encoding section the moving image signal is output, thereby controlling the moving image switching section 1001. The switch control section 1041 outputs a signal indicating the switch state to the encoding control section 1042 and the multiplexing control section 1043 at the same time.

Based on the signal supplied from the switch control section 1041, as in Example 9, the encoding control section 1042 determines which of the moving image signal or partial moving image signal is used for each encoding section 1002-1 to 1002-N. The signal is encoded. As in Example 9, the multiplexing control section 1043 controls the multiplexing section 1003 to multiplex the input moving image signal based on the signal supplied from the switch control section 1041.

(Example 11)

Fig. 23 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 11 of the present invention.

In the apparatus of Example 11, a screen dividing section 1005 is provided upstream of the moving picture switching section 1001 shown in FIG. The screen dividing section 1005 divides an input moving image signal into a plurality of segment screens, generates partial moving image signals corresponding to the segment screens, and outputs the partial moving image signal to the moving image switching section 1001. Therefore, in this example, a high-precision moving picture can be directly input into the moving picture encoding and multiplexing device without the preceding screen division process.

Therefore, in this example, a plurality of individual moving picture signals and a plurality of partial moving picture signals generated by dividing a high-precision moving picture signal are simultaneously encoded and multiplexed. In the apparatus of this example, encoding and multiplexing can be performed even when all input signals are individual moving image signals or they are partial moving image signals generated by screen division.

In this example, a screen split is provided. In addition, it is possible to provide a plurality of screen division sections in parallel, and simultaneously receive a plurality of moving image signals for screen division. In addition, instead of providing a screen division for performing screen division on an input moving picture, a moving picture time division is used which divides a moving picture signal every several frames and outputs the result to the moving picture switching part.

(Example 12)

Fig. 24 is a block diagram of a moving picture encoding and multiplexing apparatus according to Example 12 of the present invention.

In the apparatus of Example 12, a plurality of accumulation sections 1006-1 to 1006-N, an adder 1007, and a frame number calculation section 1008 are additionally provided in the apparatus of Example 11 shown in FIG.

The accumulation sections 1006-1 to 1006-N accumulate the moving image signals output from the encoding sections 1002-1 to 1002-N. The adder 1007 adds the number of bits of the motion picture encoded signal accumulated in each of the accumulation sections 1006-1 to 1006-N. Frame bit number calculation section 1008 calculates the number of bits per frame of the moving image coded signals accumulated in accumulation sections 1006-1 to 1006-N.

In the apparatus having the above-mentioned constitution, the moving image encoded signals output from the encoding sections 1002-1 to 1002-N are temporarily accumulated in the accumulation sections 1006-1 to 1006-N. The number of bits of the motion code signal in each of the accumulation sections 1006 - 1 to 1006 -N is added by the adder 1007 , and the added number of bits is output to the code control section 1042 . The encoding control section 1042 calculates a quantization width based on the number of bits received from the adder 1007 .

In this example, according to the fixed output bit rate in the multiplexing section 1003, the quantization width is set large when the number of bits is large, and small when the number of bits is small. The encoding control section 1042 outputs a quantization width obtained in this way to each of the encoding sections 1002-1 to 1002-N. Encoding sections 1002-1 to 1002-N encode the received moving image signal according to the quantization width.

Therefore, in this example, according to the output bit rate in the multiplexing section, although the output bit rate is fixed, the output signal can be encoded. In addition, since the same quantization width can be used for compression/encoding of input moving pictures with different precisions, the quality of a plurality of output moving pictures can always be kept uniform. In addition, since the quantization control is carried out based on the sum of the bits of the moving picture coded signals accumulated in each accumulation section 1006-1 to 1006-N, the output bit rate can be effectively used.

In this example, the quantization widths for encoding sections 1002-1 to 1002-N are set. In addition, other values such as bit rate may also be set.

Therefore, in this example, a plurality of partial moving picture signals obtained by dividing a moving picture by screen division can be encoded and multiplexed simultaneously.

In Examples 9 to 12, a moving picture switching section is provided upstream of the encoding section. In addition, two or more moving picture switching sections may also be provided in parallel. Also, only moving picture signals are coded and multiplexed in these examples. In addition, the device may be configured such that other signals such as audio signals are also multiplexed.

(Example 13)

An image transmission apparatus according to Example 13 of the present invention will be described with reference to FIGS. 25 to 27. FIG.

FIG. 25 is a block diagram of an image transmission of the apparatus of Example 3. FIG. Referring to FIG. 25, the image transmission apparatus includes image encoding sections 2001, 2002, 2003, . . . and a multiplexing section 2010. The image encoding sections 2001, 2002, 2003, ... respectively include encoding subsections 2004, 2005, 2006, ... and intermediate buffers 2007, 2008, 2009, .... The multiplexing section 2010 includes an input/output control subsection 2010a and an output buffer 2010b.

The operation of the apparatus having the above-mentioned constitution will now be described.

The image encoding sections 2001, 2002, 2003, ... compress/encode respective input image signals. The encoding subsections 2004, 2005, 2006, ... of the image encoding part 2001, 2002, 2003, ... use MPEG2 encoding, that is, the input image signal is encoded to conform to ISO/IEC 13818-2 (commonly referred to as MPEG2 video) MP@ML image signal, and, each preset fixed multiplexing cycle, to each intermediate buffer 2007, 2008, 2009, ... output a desired multiplexing cycle in the next multiplexing cycle The number of encoded data to multiplex. According to the control of the input/output control subsection 2010a, the multiplexing section 2010 receives and occupies the output buffer 2010b every multiplexing cycle in the middle of the image coding section 2001, 2002, 2003, ... All coded data accumulated in buffers 2007, 2008, 2009, . . . Then, according to the order of the received coded data, the coded data is sequentially output to a transmission line at a transmission rate of the transmission line. As far as encoded data can be acquired in the output buffer during the receiving period, the reception order of encoded data from the intermediate buffer of the image encoding section is not limited. Therefore, the multiplexing section 2010 determines that a buffer of encoded data in the output buffer 2010b occupies Bmux_OC when the reception is completed, and outputs the result thereof to the image encoding sections 2001, 2002, 2003, . . . Utilize calculated Bmux_OC/R, where R represents the transfer rate in the transfer route, the encoding subsections 2004, 2005, 2006, ... can relatively estimate each Transmission delay time of encoded data. The input image is encoded using the estimated value image encoding section 2001, 2002, 2003, . . .

The image encoding sections 2001, 2002, 2003, ... compress/encode input images into MPEG2 video data. In the MPEG2 encoding, interframe predictive encoding for the image signal is carried out, and a predictive error component resulting from the predictive encoding is encoded by DCT. The number of bits of the encoded data is determined by increasing or decreasing the quantization width used in quantization of a DCT coefficient. When the quantization width is small, the number of generated bits is large, but the image is compressed and encoded to obtain an image close to the original image. Conversely, when the quantization width is large, the number of generated bits is small, but the resulting image is deteriorated due to distortion caused by a quantization error. Therefore, in order to encode the image signal so as to obtain an image with a uniform quality close to the original image, the quantization width needs to be controlled to be as small as possible, fixed as much as possible, and have an as small as possible Variety. In this example, the image quality and the number of generated bits are controlled according to the quantization width. The operation of this encoding part is based on known techniques except for quantization width control (see, for example, "Latest MPEG Textbook", Ascii publishing, 1995, PP103-105). Therefore, the control of the quantization width will be described in detail below.

Fig. 26 shows a configuration of an encoding subsection of the image encoding section. Referring to FIG. 26, the encoding subsection includes a basic encoding section 2011, a one-bit counter 2012, a buffer 2013 for temporarily storing encoded data, and a control section 2014. The basic coding section includes a difference processor 2011a, a DCT processor 2011b, a quantization processor 2011c, a variable length coder 2011d, and an inter-frame predictor 2011e.

The basic encoding section 2011 encodes an image signal into the MPEG2 video format. In more detail, the inter-frame predictor 2011e determines a predicted value, the difference processor 2011a calculates a predicted difference of the predicted value, the DCT processor 2011b processes the predicted error using DCT, and the quantization processor 2011c quantizes The value processed by DCT, and the variable length encoder 2011d encodes the quantized value. The quantization processor 2011c receives a quantization value from the control section 2014 . The input image signal is quantized and encoded according to the quantization width. The resulting encoded data is transferred to the buffer 2013 . The bit counter 2012 counts the number of bits of the encoded data, and transmits the generated number of bits to the control section 2014 every multiplexing cycle. The control section 2014 determines the quantization width according to the buffer occupancy Bmux_OC transmitted from the multiplexing section 2010 in the output buffer 2010b and the generated number of bits transmitted from the bit counter 2012, and the quantization width Provided to the basic encoding part 2011. The control section also calculates the number of bits to be transferred to the corresponding intermediate buffer of the image coding section, and supplies the number of bits to the buffer 2013. The buffer 2013 outputs the encoded data transferred from the basic encoding section 2011 to the corresponding intermediate buffer according to the number of bits specified by the control section 2014.

FIG. 27 shows a flowchart for calculating the quantization width by the control section 2014 of the encoding subsection. Referring to FIG. 27, at step 2201 the process begins. In step 2202, the reference quantization width q_st, the code subpart buffer occupation Bei_oc, the intermediate buffer occupation Bmi_oc and the parameter Di are initialized. In step 2203, it is judged whether the encoding has been completed. If completed, in step 2204 the process is aborted. If not, the quantization width is calculated in step 2205 . In step 2206 the resulting number of bits is checked. In step 2207 the calculated number of bits is output. In step 2208 the output buffer occupancy Bmux_oc is received, and in step 2209 the parameter Di is calculated. Then, processing to prevent overflow in the output buffer (step 2210) and the encoding skip (step 2211) are performed.

The above operations are carried out for all encoding subsections shown in FIG. 25 . In the following, the operation of the first encoding subsection will be described in detail as an example.

After initialization (step 2202), the control section 2104 uses the buffer size Bdec_Size of a hypothetical decoding device set at the receiving side to reproduce the data encoded by the i-th encoding subsection using the following equation (8) to calculate Quantization width for each multiplexing cycle (step 2205) until the encoding is completed (step 2203).

qi＝(Bdec_size/(Bdec_size-Di))·q_st (8)

Wherein q_st represents the reference quantization width to be quantized at the time of initialization so that a desired image quality can be obtained. Di represents a variable that changes every multiplexing cycle. The calculation of this variable will be described below. The encoding subsection outputs encoded data to the intermediate buffer in the image encoding section every multiplexing cycle. In more detail, the control section 2014 detects the number of bits generated from the bit counter 2012 in a multiplexing cycle. When the sum of the generated number of bits and a non-output remaining number of bits REMi is less than the size of the intermediate buffer, the control part 2014 sends an instruction output Bmi to the buffer 2013, indicating all encoded data (that is, REMi and the generated The sum of the number of bits) will be output to the intermediate buffer within the range not exceeding the upper limit specified in a standard (ie, 15 Mbps below MP@ML). When the sum is greater than the size of the intermediate buffer, the control section 2014 sends an instruction output amount Bmi to the buffer 2013, specifying that an amount equal to the capacity of the intermediate buffer will be output, while keeping the remainder in the buffer 2013 as an unfinished Output the remaining data (steps 2206 and 2207).

After the output buffer occupancy Bmux_oc can be calculated and the encoded data is multiplexed for one multiplexing cycle received at the receiver side, when the amount of encoded data encoded by the encoding subsection but not reproduced is greater than that of the decoding For the buffer size Bdec_size of the device, the command output amount Bmi is determined such that the value obtained by subtracting the buffer size Bdec_size from the amount of coded data that has not been reproduced is equal to the remaining number of unoutput data bits REMi. In other words, the buffer of the decoding device is prevented from overflowing by keeping the amount of data to the receiver side equal to the buffer size Bdec-size. More specifically, the upper limit of the amount of encoded data is set to a value obtained by subtracting the buffer occupancy from the buffer size of the decoding device. For example, the above-mentioned amount of non-reproduced encoded data is determined as follows. In this case, the weight time of the designated decoding device is to use a time stamp added to the encoded data so that when the buffer occupancy Bmux_oc in the output buffer 2010b becomes equal to the buffer size Bmux_size In the case where the encoded data is reproduced by the decoding means at a delay time of ○, the number of bits bT generated in each encoding subsection during the time (Bmux_size-Bmux_oc)/R is calculated and totaled. Then, use formula (9) to calculate the number of bits of unoutput data:

REMi＝REMi+bT-Bmi (9)

where bT is the number of bits generated in one multiplexing cycle.

Afterwards, the control section 2104 receives the current data possession Bmux_oc of the output buffer 2010b of the multiplexing section 2010, and calculates the parameter Di with the formula (10) on the premise of the formula (11) (steps 2208 and 2209).

Di＝Bdec_size*Bmux_oc/Bmux_size+f(REMi) (10)

f(REMi)＝REMi (11)

The quantization width is calculated from the parameter weight Di thus obtained using equation (8).

In order to prevent the value of subtracting Di from the decoder buffer size (Bdec_size-Di) from becoming 0, the control section 2014 instructs the basic encoding section 2011 to perform skipping when (Bdec_size-Di) is equal to or smaller than the threshold K ( Steps 2210 and 2211).

The multiplexing section 2010 receives the encoded data stored in the intermediate buffer of each image encoding section in each multiplexing cycle, according to ISO/IEC 13818-1 (commonly referred to as MPEG2 TS stream) The encoded data is multiplexed into a data packet, and the data packet is output to the transmission line. At this time, the reproduction time indicated by the time stamp for specifying the reproduction timing of each picture to the encoding apparatus is determined from the buffer occupancy Bmux_oc of the output buffer 2010b. That is, reproduction of the encoded data starts at delay time 0 when the output buffer 2010b becomes full. Using this time as a reference, the recurring time is determined and the time stamp designating the recurring time is added to each transmitted data packet.

With the above configuration, the initial transfer start time in the multiplexing section 2010, the output buffer size Bmux_size of the multiplexing section 2010, etc. actually depend on the image coding section N connected to the multiplexing section 2010. The number, transmission rate, etc. vary. However, in this example, when Bmux_size=N×Bdec_size is set, it is assumed that the encoded data is multiplexed immediately after encoding and arrives at the decoding device when the output buffer occupancy Bmux_oc is 0 and it is assumed that X images (ie, there is a possibility that this data corresponds to the maximum X images accumulated in the decoder) to reproduce the coded data afterwards. Using the case where the output buffer occupancy Bmux_oc is 0 as a reference, when the output buffer occupancy Bmux_oc is an arbitrary value, the relative delay time of the multiplexed encoded data is calculated according to Bmux_oc/R. Therefore, when the data arrives at the decoding device, the data size of the image held in the decoding device is X·(Bmux_size-Bmux_oc(t))/Bmux_size (when the non-input data size is 0). In this case, the encoded data received by the receiver side does not underflow.

More specifically, first, when REMi=0, the following formula (12) is obtained from formulas (8) and (9).

qi＝(Bmux_size/(Bmux_Size-Bmux_oc))·q_st (12)

It can be seen from equation (12) that the closer the output buffer occupation Bmux_oc is to the output buffer size Bmux_size, the larger the quantization width will be. Therefore, the number of bits generated is compressed to prevent underflow from occurring. Even if an extremely complex image is output, Bmux_oc is always controlled to be smaller than Bmux_size by skipping step 2211 in FIG. 27 . This prevents underflow from occurring.

When REMi≠0, the following formula (13) is obtained.

qi＝q_st/(1-(Bmux_oc/Bmux_size)-REMi/Bdec-Size) (13)

It can be seen from equation (13) that the image coding part is controlled so that REMi is smaller than Bdec_size*(Bmux_size-Bmux_oc)/Bmux_size. Even if an extremely complex image is output, the image encoding part is controlled by reducing the number of generated bits by skipping in step 2211 in FIG. 27 so that REMi is smaller than Bdec_size*(Bmux_size-Bmux_oc)/Bmux-Size. Even if all other image coding sections need to multiplex coded data of the maximum number of bits, the coded data can be multiplexed at least at an average rate obtained by dividing the transfer rate by the number of image coding sections. Since Bdec_Size×(Bmux_size−Bmux_oc)/Bmux_size corresponding to the number of bits of encoded data can be transmitted at an average rate within time (Bmux_size−Bmux−OC)/R, underflow does not occur on the receiver side.

Since coded data exceeding the buffer capacity of the decoding device is left as non-output data, overflow does not occur and therefore all coded data coded by the image coding section still maintains repetition.

With the above constitution, the image encoding section determines the quantization width according to the same standard. Therefore, the quantization width used in all picture coding sections is also independent of the input picture, and thus the input picture can be coded giving substantially the same picture quality. Even if a scene change is included in a continuous input image, since the quantization width is controlled according to the number of bits generated in all image coding sections, the number of bits is specified according to the increase in its complexity The image quality at the scene change will not be greatly reduced, while the impact of the bit increase and the change in image quality remain small. Each image coding section controls the number of bits generated and the number of bits of coded data to be multiplexed so that the coded data does not overflow and underflow on the receiver side. Therefore, the multiplexing section 2010 is only suitable for multiplexing the received regular coded data, and does not need to perform information processing on the entire image as conventionally required. Therefore, since the processing is distributed, the device is easy to construct.

In this example, the quantization width is calculated by equation (8). Additional formulas can be utilized in terms of maintaining Bdec_size-Di>0.

Another formula can be used for f(REMi)=REMi of equation (11). The decision characteristic used to determine the quantization width and f(REMi) for a given picture code portion can be intentionally varied in order to maintain a high picture value level for a particular picture signal. For example, f(REMi) may be REMi/M when the capacity of the intermediate buffer of the j-th encoding section is made larger than M times the capacity of other intermediate buffers. In order to do this, the amount of non-output coded data can be increased to an amount close to M times the allowed REMi in this example. Under normal circumstances, when an image is more complex than the average of all input images, when the encoded data arriving at the decoding device at the Bmux_oc time is multiplexed, it corresponds to the encoded X (Bmux_size-Bmux_oc( t))/Bmux_size The data size of the image is limited to Bdec_size+REMi. Therefore, for a complex image that is difficult to encode, having more unoutput data may allocate relatively more bits.

In this example, the shorter the multiplexing period, the more precise the control. However, for the multiplexing period, several microgroups or several parts (several lines of the image signal) of MPEG2 are suitable. A long period of one image or the like may also be used.

In this example, the data transfer method of the encoding method is the same for all image encoding sections. However, under a picture quality control different from the standard picture quality control, a specific basic coding part can be processed. For example, in terms of the function that the quantization width increases as Di increases, any value can be used as the judgment function for the quantization width in equation (8). The image quality can be kept uniform by reducing the rate of change of the quantization width. Also, when all the picture coding sections operate as described above, in a specific picture coding section, the quantization width can be fixed to make the picture quality uniform, or in the picture coding section the quantization width can be fixed. Data can be encoded at a fixed rate to generate bits. In this case, overall, the number of bits generated tends to decrease as Di increases. Therefore, the overall bit rate can be within the bit rate range on the transmission line.

(Example 14)

An image transmission apparatus according to Example 14 of the present invention will be described below with reference to FIGS. 28 to 34. FIG.

Fig. 28 schematically shows the image transmission apparatus of Example 14. Referring to FIG. 28, the image transmission apparatus includes a bit rate management node 2015 for controlling the bit rate on a LAN (Local Area Network), and nodes 2016 to 2016 connected to the LAN having an image encoding section. 2019. Each of the nodes 2016 to 2019 includes an image reproduction section and an image encoding section. The picture coding part codes the picture taking into account the buffer size Bdec-size and the buffer occupancy of the decoding means of the picture reproduction part of an object node.

With the above constitution, each image encoding section disposed on a node connected to another transmits encoded data to another image encoding section on the LAN as a packet via the LAN. The packet includes a header information and encoded data required for transmission.

Fig. 29 shows a header format. The header information includes information required for the correct transfer of the encoded data, ie, presumably known information required on the LAN. In this example, description of such information is omitted, and only a sender address, a receiver address, a picture packet flag, and a transfer cycle number are described. In this network, the path of a transmission line is extended between each adjacent node, forming a loop. The bit rate management node 2015 controls the generation of packets. That is, the bit rate management node 2015 acquires the bit rate based on the request from the node and the information generated for transmission.

Fig. 30 illustrates the communication method between nodes. For example, in communication between nodes 2016 and 2019, the bit rate management node 2015 generates a packet with a sender address of 1 and a receiver address of 4, and sends an empty packet to the network Output the packet regularly. Each node transmits the packet to neighboring nodes via a packet that does not refer to itself. In receiving the sender address packet with 1, the node 2016 puts the encoded data that the node 2016 wishes to transmit in the packet, and transmits the packet to the next node 2017. Nodes 2017 and 2018 send this information packet to the next node. Node 2019 detects its own address, address 4, as the receiver address in the packet, receives the encoded data in the packet, and then releases the packet.

Transmission of encoded data representing images will be described below with reference to FIGS. 31 to 36. FIG.

The bit rate management node 2015 obtains the complete bit rate for transmitted encoded data in the network. For example, assume that the number of nodes transmitting encoded data in the network is 4 and that the average bit rate for each node is approximately 3 Mbps. In this case, the bit rate management node 2015 gives a bit rate of 12 Mbps and outputs packets for image transmission corresponding to 12 Mbps. Each respective packet for image transmission is provided with a multiplex cycle number which is updated every multiplex cycle. Each node recognizes the multiplexing cycle number of each received packet and outputs the packet during the multiplexing period indicated by the multiplexing cycle number. Before the packet is output, each node notifies the number of output packets at each multiplexing cycle.

Fig. 31 is an example of the transfer of the packet for notifying the number of packets. In this packet, the two sender and receiver addresses are ff and the image packet flag is 1, so that each node recognizes the packet for notifying the number of packets. Upon receiving such a packet, each node writes the number of packets transmitted through the node during the multiplexing period corresponding to the indicated number of multiplexing cycles. The bit rate management node 2015 reads the packet that is circulated through the network for notifying the number of packets, and outputs the same multiplexing cycle number for the number equal to the number of points of the notified packet The packet of image transmission.

Fig. 32 is an example of transmission of coded data using the packet for image transmission. The bit rate management node 2015 generates a packet for image transmission having an information flag bit of 1 and a certain number of multiplexing cycles, and outputs the packet on the network. This packet is available to any node that has image transfer notifications. Therefore, any node wishing to transmit its coded data in the specified multiplexing cycle writes its own address and receiver address in the header of the information packet, puts the coded data in the information packet, And transmit the information packet to the node of the device. In the example shown in FIG. 32, transmission from node 2016 to node 2019 is shown. Node 2019 receives the packet, resets the data in the packet, and outputs the packet in the network. When the node 2016 located closest to the bit rate management node 2015 has transmitted all the information that the node 2016 intends to transmit between the specified multiplexing cycles, the following multiplexing cycles have the same Number of information packets arrive at the next node through node 2016. As explained above for node 2015, between specified multiplexing periods, node 2017 located closer to the bit rate management node 2015 places its transmitted encoded data in in the packet and transmit the packet.

In this way, each node receives a number of packets equal to the number of packets advertised by the node, within which its The encoded data is transmitted and the packet is output in the network in such a manner that the node closer to the bit rate management node 2015 does so earlier. Since the number of packets having the same multiplexing cycle generated by the bit rate management node 2015 is equal to the total number of packets notified by these nodes, the packets are neither short nor long. Therefore, packets for the next multiplexing cycle are generated based on the notification by these nodes.

The multiplexing cycle used in this example corresponds to the multiplexing cycle described in Example 11, and is actually generated by calculating until the bit rate management node 2015 according to the total number of notified packets Determined by the delay time required to have the same multiplex number.

For example, assuming that the multiplexing period is T and the amount of delayed data after TXK is DD(KT), the amount of delayed data after TX(K+1), that is, DD((K+1)T) is given by Calculated by the following formula (14).

DD((K+1)T)＝DD(KT)+Pn P_DATA-RT (14)

Where Pn represents the total number of packets notified at this time, P_DATA represents the data amount of each packet, and R is the bit rate for image transmission given by the bit rate management node 2015.

Then, calculate the virtual output buffer size (hereinafter referred to as the multiplexing virtual buffer size) Bmux_size of all encoding parts according to the following formula (15), and calculate the relative delay energy delay_mux by the following formula (16) . The resulting multiplexed virtual buffer size and relative delay energy according to the format shown in Fig. 29 is put into the packet for image transfer.

Bmux_Size＝N×Bdec_size (15)

delay_mux＝Bmux_size-(delay_init×R)+DD (16)

Among them, delay-init represents the delay time from the arrival of the transmitted data when the delay data amount DD is 0 until the data is reproduced, and this delay time is shared by the nodes in the network. The calculated value is put into the header of the packet generated immediately after the calculation, and transmitted.

FIG. 33 shows the transfer of packets generated by the bit rate management node 2015.

A packet for notifying the number of packets is generated every predetermined period. Packets for image transmission are generated from the number notified for notifying the number of packets in the packet according to the rate R given by the bit rate management node 2015 for transmitting all encoded data representing images. However, unlike the case when the delay data amount increases to equal to delay_init×R during a multiplexing cycle, the generation of the packet is suspended during this multiplexing cycle and even if the number of generated packets is short The number of multiplexing cycles is also updated according to the number of notified packets. In the case when the generation of the notified number of packets is completed within one multiplexing cycle, packets with the updated number of multiplexing cycles are generated in the next multiplexing cycle. Therefore, coded data of the number of packets notified by these nodes is transmitted so that the multiplexing delay number DD does not exceed delay_init×R.

Fig. 34 shows the configuration of each node. Referring to Fig. 34, this node comprises a packet processing section 2020, a control section 2021, an intermediate buffer 2022, and an encoding section 2023, and an image reproduction section 2024 for processing the input/output of the packet .

The packet processing section 2020 receives a transmitted packet through a network and detects the header of the packet. When the receiver address of the header is its own address, the packet processing section 2020 passes the contents of the packet to the control section 2021 and outputs the information immediately after resetting the sender address and receiver address of the packet Bag. When the transferred data in the packet is encoded data representing an image, the control section 2021 transfers the encoded data to the image and image reproduction section 2024 for attaching the image.

When the address of the sender of the received packet is its own address, the packet processing section 2020 writes encoded data in the packet according to the information from the control section 2021 and the intermediate buffer 2022, and outputs the resultant information package.

When the header information indicates that the packet is for notifying the number of packets, together with the number of multiplexing cycles in the packet, the packet processing section 2020 transmits to the control section 2021 a message indicating that the packet is Heartbeat to notify the number of packets. The control section 2021 transmits to the packet processing section 2020 a notified packet number indicating a multiplexing cycle. The notified packet number for this multiplexing period corresponds to the number of data transmitted from the encoding section 2023 to the intermediate buffer 2022 during the specified multiplexing period. This number of data is calculated by the intermediate buffer 2022, and the calculated value is determined as a value notified from the control section 2021. The packet processing section 2020 writes the notified value in the packet and outputs the resulting packet to the network.

When the header information of the packet indicates that the packet is for image transmission, the packet processing section 2020 sends to the control section 2021 a detection signal indicating that the packet is for image transmission. The control section 2021 then reads the number of multiplexing cycles of the packet. If the number of packets that have been output during the indicated multiplexing cycle is shorter than the notified number of packets, the control section 2021 instructs the packet processing section 2020 to write the code representing the image along with the receiver address. data. The packet processing section 2020 writes its own address as the sender address and specifies the receiver address as the target address, which fetches coded data corresponding to a message from the intermediate buffer 2022, writes the Encodes the data and outputs the resulting packets into the network. After outputting the notification packet number during this multiplexing cycle, the control section 2021 updates the multiplexing cycle number. The control section 2021 also sends an output instruction signal to the packet processing section 2020 to allow output of a packet followed by the old number of multiplexing cycles without any processing before updating.

When receiving a packet for image transmission, the control section 2021 receives the multiplexing virtual buffer size Bmux_size and the relative delay parameter delay_mux of the packet from the packet processing section 2020, and transmits these values to the encoder Section 2023. These values are always checked independently of the transfer timing.

The operation of the encoding section 2023 will be described below. The operation of this encoding section 2023 is basically the same as that of the encoding subsection in Example 13 shown in FIG. 26 .

In this example, the control section 2014 shown in FIG. 26 receives the multiplexed virtual buffer size Bmux_size and the relative delay parameter delay_max from the control section 2021 in FIG. 34, and performs the same operation as described in Example 13. , use these values to replace the buffer size Bmux_size and buffer occupancy Bmux_oc of the multiplexing section 2010 in FIG. 13 (see FIG. 25 ). In this example, since the transfer rate R set by the bit rate management node 2015 is equivalent to the transfer rate in the multiplexing section 2010 in Example 13, it is divided between the encoding sections of the node. The coded data transmitted and received is transmitted with a timing at which the coded data can be reproduced without overflow or underflow on the receiver side.

Therefore, according to this example, images can be transmitted over a network that keeps the image quality substantially uniform. 35A and 35B show a comparison between conventional transmission of encoded data with a variable bit rate over a network and transmission of encoded data at the bit rate used in this example. Conventionally, the bit rate for the coded data coded according to the complexity of the picture is acquired on the transmission line. In this case, as shown in FIG. 35A, for example, when three blocks of coded data are transmitted, all bit rates on the transmission line are not effectively used. However, in this example, as shown in FIG. 35B, since all bit rates on the transmission line are always used for encoding, the bit rates are effectively used. Because data is encoded using a large number of bits, high-quality images can be reproduced. When the image quality is not high, more images can be transmitted on the network.

In this example, the network connects the nodes in a ring shape as described. The present invention can also be used for networks of other shapes or networks with different transmission protocols until the bit rate specified for an image on the transmission line is obtained, the amount of delay in the transmission is provided to each node, And the encoded data generated by the encoding section can be transmitted.

The delay time represented by delay_init can be set arbitrarily within the range of Bmux_size-(delay-initxR)≥0. In particular, when Bmux_size-(delay_init=R)=0 is satisfied, more efficient variable-length encoding is possible. As in Example 13, only a specific fix can be encoded to obtain high quality by changing the setting of a specific encoding section. Also, in this example, at least the average bit rate obtained by dividing the entire bit rate of the input data can be obtained per multiplexing cycle by multiplexing the maximum allowable data volume, so it is possible to transmit at an average rate coded video signal.

In this example, the case where the number of images to be transferred is not changed is described. Additionally, the number of images may vary during this transfer. In this case, Bmux_xize and delay_mux will be updated. In the case where the relative delay parameter delay_mux becomes pages due to changes in the number of pictures during the transfer, delay_mux will be updated to a value smaller than the value obtained by equation (16) before appending a new picture , reduce the delay_mux to such an extent that the delay_mux will not be paged by appending a new image.

As shown in Fig. 36, the nodes in this example can be implemented by an ordinary computer by inserting the algorithm described as a software in this example. The present invention is applicable to encoding of an image signal obtained from a video camera or an image signal in a memory such as a hard disk of an ordinary computer.

Formula 1

Formula 2 P3i=min (occupancy of the encoder buffer, free capacity of the decoder buffer)-P2i

Formula 3 D = output digits - ∑P2i

Formula 4 Number of multiplexed bits = P2i+P3i

Formula 5

Formula 6

Formula 7

Formula 8 qi＝(Bdec_size/(Bdec_size-Di))·q_st

Formula 9 REMi＝REMi+bT-Bmi

Formula 10 Di＝Bdec_size*Bmux_oc/Bmux_size+f(REMi)

Formula 11 f(REMi)＝REMi

Formula 12 qi=(Bmux_size/(Bmux_size-Bmux_oc))·q_st

Formula 13 qi＝q_st/(1-(Bmux_oc/Bmux_size)-REMi/Bdec_size)

Formula 14 DD((k+t)T)＝DD(kT)+Pn·P_DATA-RT

Formula 15 Bmux_size＝N×Bdec_size

Formula 16 delay_mux＝Bmux_size-(delay_init×R)+DD

Claims (40)

1. A moving image encoding device for encoding a plurality of moving image signals, comprising: a plurality of encoding sections for encoding a plurality of moving image signals, wherein each encoding section is provided with a quantization subsection for quantizing the plurality of moving image signals by a preset quantization width; and a quantization control section for simultaneously controlling the quantization widths used by the quantization subsections; and further comprising a plurality of accumulation sections for temporarily accumulating a plurality of moving image encoded signals output from the plurality of encoding sections, Wherein, the quantization control section determines the quantization width used by the quantization sub-section based on the sum of the bits of the plurality of moving image coded signals accumulated in the accumulation section. 2. The moving picture coding apparatus according to claim 1, wherein the quantization widths used by the quantization subsections are the same. 3. The moving picture coding apparatus according to claim 1, further comprising a screen separating section arranged upstream of the plurality of encoding sections for separating an input moving picture signal into a plurality of moving picture signals. 4. A moving image multiplexing device for multiplexing a plurality of encoded moving image signals obtained by encoding a plurality of moving image signals, comprising: a plurality of accumulating parts for temporarily accumulating a plurality of moving image coding signals; a multiplexing section for multiplexing a plurality of motion picture coded signals accumulated in a plurality of accumulation sections and outputting the multiplexed signal at a fixed bit rate; a decoding delay time calculation section for calculating a period of time from when a specific bit of each of the plurality of moving image coded signals is input until when the specific bit is decoded by the decoding means; a frame bit number calculation section for calculating the bit number of a frame including a specific bit of each of the plurality of moving image encoded signals accumulated in the plurality of accumulation sections; and A multiplexing control section, used for calculating a plurality of first input bit numbers according to the calculation results of the decoding delay time calculation section and the frame bit number calculation section, and the plurality of first input bit numbers are obtained from a plurality of The accumulating section inputs the number of bits of a plurality of moving picture coded signals input to the multiplexing section. 5. The moving picture multiplexing apparatus according to claim 4, wherein the multiplexing control section comprises: A first parameter generation subsection is used to determine a plurality of first parameters, wherein each parameter in the plurality of first parameters is the number of bits calculated by the frame bit calculation part and the number calculated by the decoding delay time calculation part The ratio of time to calculate; a second parameter generating subsection for determining a second parameter that is the sum of a plurality of first parameters; a third parameter generating subsection for determining a plurality of third parameters, wherein each of the plurality of third parameters is a plurality of first a ratio of each of a parameter to said second parameter; and A multiplexing bit calculation part, used to calculate a plurality of first input bit numbers according to a plurality of third parameters. 6. The motion picture multiplexing device according to claim 5, further comprising; an accumulation bit calculation section for calculating the number of bits of the plurality of moving image coding signals accumulated in the plurality of accumulation sections; and an unoccupied capacity calculation section for calculating an unoccupied capacity of an input buffer of a decoding device immediately before specific bits of each of a plurality of moving image coded signals are input to the decoding device, Wherein the multiplexing number of digits calculation part sets each of the plurality of second input digits as follows: (1) The first input digit determined according to the third parameter, (2) the number of bits of the moving image coded signal calculated by the accumulated number of bits calculation section, and (3) the unoccupied capacity of the input buffer of the decoding device calculated by the unoccupied capacity calculation section, a minimum value among , and The multiplexing bit number calculation section controls the number of bits of the moving picture coded signal sent from the accumulation section to the multiplexing section based on the plurality of second input bit numbers. 7. The moving picture multiplexing apparatus according to claim 6, wherein when the plurality of second input bit numbers calculated by the multiplex bit number calculation section are different from the corresponding first input bit numbers, the plurality of second input bit numbers The multiplexing bit calculation section increases the second input bit. 8. A moving image multiplexing method for multiplexing a plurality of moving image coded signals after being accumulated in a plurality of accumulation parts, comprising the steps of: (1) calculating a period of time from when a specific bit of each of a plurality of moving image coded signals is input to a decoding device having an input buffer until when the specific bit is decoded, (2) calculating the number of bits of one frame including a specific bit of each of the plurality of moving image coded signals accumulated in the plurality of accumulation sections; and (3) A plurality of bits to be multiplexed from a plurality of moving image coded signals is determined based on the results calculated from steps (1) and (2). 9. The moving image multiplexing method according to claim 8, wherein step (3) comprises the steps of: (4) determine a plurality of first parameters, each of the plurality of first parameters is the ratio of the number of digits obtained in step (2) to the time obtained in step (1); (5) determine the second parameter, the second parameter is the sum of multiple first parameters; (6) determining a plurality of third parameters, each of which is a ratio of each of the plurality of first parameters for the plurality of motion image encoding signals to the second parameter; (7) A plurality of multiplexing bit numbers are determined according to a plurality of third parameters. 10. The motion picture multiplexing method according to claim 9, wherein an input multiplexing bit number for multiplexing each of a plurality of motion picture coded signals is set as: (1) according to the number of multiplexing bits determined by the third parameter, (2) the number of bits of the moving picture coded signal accumulated in the accumulation section, and (3) an unoccupied capacity of the input buffer of the decoding device, a minimum value among , and Each bit that is multiplexed is controlled by the number of input multiplexing bits. 11. The moving picture multiplexing method according to claim 10, wherein when at least one input multiplexing bit is different from the corresponding multiplexing bit, the input multiplexing bit is increased. 12. A moving picture multiplexing method for multiplexing a plurality of moving picture coded signals, a plurality of multiplexing bits corresponding to a plurality of moving picture coded signals which are preset to zero , the method includes the steps: (1) calculating the number of frame bits of a frame including a specific bit of each of a plurality of moving image coded signals; (2) calculating a period of time until each of the plurality of moving image coded signals is decoded; (3) Select the moving image coded signal whose time period obtained in step (2) is slightly shorter than a preset time value; (4) determine a plurality of main multiplexed bit numbers according to the frame bit numbers of the motion picture coded signal selected in step (3); (5) calculating the sum of a plurality of first multiplexing digits; (6) adding the plurality of first multiplexing bits to the plurality of multiplexing bits; and (7) update the preset time value, Wherein steps (1) to (7) are repeated, and, when the sum of a plurality of first multiplexed digits is greater than a preset digit number, then determine a plurality of times multiplexed digits and add to multiple multiplexing bits to multiplex the moving image coded signal of a plurality of multiplexing bits. 13. A motion picture multiplexing method for multiplexing a plurality of motion picture coded signals, comprising the steps of: A plurality of minimum transfer amounts required for preventing underflow of a plurality of input buffers of the decoding device and determining up to a plurality of motions are determined based on the number of bits of frames of a plurality of moving image coded signals not transferred to the decoding device. The time required until a frame of an image coded signal is decoded; determining a lower limit of the plurality of transfer bits based on a plurality of minimum transfer amounts; and The lower limit of the plurality of transmission bit numbers is multiplexed before multiplexing the plurality of motion picture coded signals. 14. A moving image encoding device for encoding a plurality of moving image signals, comprising: A plurality of encoding sections for encoding a plurality of moving image signals, each of the plurality of encoding sections includes a quantization subsection for quantizing the moving image signal with a preset parameter and a quantization subsection for quantizing the moving image signal with a preset parameter. The amount of padding is used to stuff the stuffing sub-section of the quantized motion picture signal, a quantization control section for controlling parameters also used by the quantization subsection; a plurality of accumulating sections for temporarily accumulating a plurality of moving image encoded signals output from the plurality of encoding sections; a multiplexing section for multiplexing a plurality of motion picture coded signals accumulated in a plurality of accumulation sections and outputting the multiplexed signal at a fixed bit rate; a stuffing control section for controlling the filling volume used by the stuffing subsection; and a virtual unoccupied capacity calculation section for calculating the input buffers of the plurality of decoding devices after skipping the preset time period when the plurality of motion picture coded signals are not input to the input buffer for the preset time period an unoccupied capacity of each of the wherein the quantization control section determines parameters used by the quantization subsection based on the sum of the number of bits of a plurality of moving image coded signals, and According to the sum of the number of bits of the moving picture coded signal accumulated in a plurality of accumulation parts, the fixed bit rate, and the number of bits of each moving picture coded signal accumulated in the accumulation part and the virtual unoccupied Each stuffing subsection is controlled by the difference between the unoccupied capacities of the input buffers obtained by the capacity calculation section. 15. A moving picture coding method for simultaneously coding a plurality of moving picture signals using preset quantization widths based on a plurality of moving picture signals generated by encoding a plurality of moving picture signals the difference between the sum of the number of bits of the motion picture coded signal and the output number of bits of each of the plurality of motion picture coded signals output after temporal accumulation, wherein a plurality of coded buffer occupations of temporally accumulated motion picture coded signals are calculated, calculating, for a preset time period, an unoccupied capacity of each of the input buffers of the plurality of decoding means after skipping a preset time period when the motion picture coded signal is not input to the input buffer; and For motion picture coded signals whose buffer difference obtained by subtracting the unoccupied capacity of the input buffer from the coded buffer occupancy is less than a preset value, it will be added by adding a value greater than a preset value that does not affect the encoding. The one-digit signal obtained by the buffer difference value is distributed to the motion image coding signal whose buffer difference value is smaller than a preset value to realize stuffing. 16. The moving picture coding method according to claim 15, wherein the preset value is zero. 17. A moving picture encoding device for encoding a plurality of moving picture signals, comprising: a plurality of encoding sections for encoding a plurality of moving image signals, each of the plurality of encoding sections includes a quantization subsection for quantizing the moving image signal with a preset parameter and a quantization subsection for quantizing the moving image signal with a preset parameter The stuffing amount to stuff the stuffing sub-part of the quantized motion picture signal; a quantization control section for controlling parameters simultaneously used by the quantization subsections; a plurality of accumulating sections for temporarily accumulating a plurality of moving image encoded signals output from the plurality of encoding sections; A multiplexing section for multiplexing the plurality of motion picture coded signals accumulated in the plurality of accumulation sections and outputting the multiplexed signal at a fixed bit rate. a virtual unoccupied capacity calculation section for calculating, for a preset time period, inputs of a plurality of decoding devices after a preset time period is skipped when a plurality of moving image coded signals are not input to the input buffer the unoccupied capacity of each of the buffers, Wherein the quantization control part is based on the sum of the number of bits of the moving picture coded signal accumulated in a plurality of accumulation parts, the fixed bit rate, and the number of bits of the moving picture coded signal accumulated in the accumulation part The difference between the unoccupied capacity of the input buffer obtained by the occupied capacity calculation section determines each parameter used by the quantization subsection. 18. A moving picture coding method for simultaneously coding a plurality of moving picture signals using preset quantization widths based on a plurality of moving picture signals generated by encoding a plurality of moving picture signals The difference between the sum of the number of bits of the motion picture coded signal and the output number of each of the plurality of motion picture coded signals output after the temporary accumulation comprises the steps of: Computing multiple coded buffer occupancy of temporally accumulated motion picture coded signals, For a preset time period, calculating the unoccupied capacity of each of the input buffers of the plurality of decoding devices after skipping the preset time period when the motion picture coded signal is not input to the input buffer; and Each motion picture encoded signal is encoded using a quantization width smaller than the corresponding preset quantization width, where a buffer difference obtained by subtracting the unoccupied capacity of the input buffer from the encoding buffer occupancy is less than A preset value. 19. The moving picture coding method according to claim 18, wherein the preset value is zero. 20. A moving image encoding method for simultaneously encoding a plurality of moving image signals using preset quantization widths, wherein the preset quantization widths are based on the encoding of the plurality of moving image signals The difference between the sum of the number of bits of a plurality of moving image coded signals and the output number of each of the plurality of moving image coded signals output after temporarily accumulating comprises the steps of: A plurality of encoding buffers of the temporarily accumulated motion picture encoding signal are calculated and occupied, For a preset time period, calculating the unoccupied capacity of each of the input buffers of the plurality of decoding devices after skipping the preset time period when the motion picture coded signal is not input to the input buffer; and Each motion picture encoded signal is encoded using a quantization width greater than the corresponding preset quantization width, where a buffer difference obtained by subtracting the unoccupied capacity of the input buffer from the encoding buffer occupancy is greater than A preset value. 21. The moving picture coding method according to claim 20, wherein the preset value is zero. 22. A moving picture coding method for simultaneously coding a plurality of moving picture signals using preset quantization widths based on a plurality of moving picture signals generated by encoding a plurality of moving picture signals The difference between the sum of the number of bits of the motion picture coded signal and the output number of each of the plurality of motion picture coded signals output after the temporary accumulation comprises the steps of: A plurality of encoding buffers of the temporarily accumulated motion picture encoding signal are calculated and occupied, For a preset time period, calculating an unoccupied capacity of each of the input buffers of the plurality of decoding devices after skipping the preset time period when the motion picture coded signal is not input to the input buffer, calculating a buffer difference obtained by subtracting the unoccupied capacity of the input buffer from the coding buffer occupancy for each motion picture encoded signal, and dividing the difference between the buffer differences greater than a preset value and added to the sum of digits. 23. The moving picture coding method according to claim 22, wherein the preset value is zero. 24. A moving picture encoding and multiplexing apparatus for encoding a plurality of moving picture signals and multiplexing the encoded signals, comprising: a moving image switching section for switching an object of a plurality of input moving images; a plurality of encoding sections for encoding a plurality of input moving images output from the moving image switching section; a multiplexing section for multiplexing encoded signals output from a plurality of encoding sections; and a control section for switching the switching state of the moving picture switching section, a coding mode of the coding section, and a multiplexing mode of the multiplexing section; Wherein the control part switches the coding mode of the coding part and the multiplexing mode of the multiplexing part according to the switching state of the moving picture switching part. 25. The motion picture encoding and multiplexing device according to claim 24, wherein the control section comprises: a switch control subsection for switching the switch state of the motion picture switch section; an encoding control subsection for to switch the encoding state of the encoding part; and a multiplexing subsection for switching the multiplexing mode of the multiplexing part, and The encoding control subsection switches the encoding mode according to the switching state of the moving picture switching part, and the multiplexing control subsection switches the multiplexing mode according to the switching state of the moving picture switching part. 26. The moving picture encoding and multiplexing apparatus according to claim 24, further comprising a screen separation section upstream of the moving picture switching section. 27. The moving picture encoding and multiplexing apparatus according to claim 24, further comprising a moving picture time division section upstream of the moving picture switching section. 28. The moving picture coding and multiplexing apparatus according to claim 25, further comprising an accumulation section for temporarily accumulating the moving picture coding signal output from the coding section, Wherein the encoding control subsection determines a parameter for controlling each encoding section according to the sum of the bits of the moving image encoding signal accumulated in the accumulating section. 29. The motion picture coding and multiplexing device according to claim 24, wherein the multiplexing section comprises: a motion picture multiplexing subsection for multiplexing a plurality of motion picture codes signal and output a moving image coded signal; a channel multiplexing subsection for multiplexing a plurality of input signals input to the multiplexing section and outputting from the moving image multiplexing subsection and a switch subsection for switching objects of moving picture coded signals input to the multiplexing section between the moving picture multiplexing subsection and the channel multiplexing subsection. 30. An image transmission device comprising a plurality of image coding sections and a transmission processing section, wherein The transmission processing section multiplexes coded data specified by the image coding section for each multiplexing timing, transmits the coded data to a predetermined line, and transmits multiplexed information of the image coding section, The multiplexing information is calculated based on the amount of encoded data in the image encoding section and the amount of delay available until the encoded data is transmitted to a decoding device, and A plurality of picture encoding sections determine a quantization width used in quantization in an image process based on the multiplexing information sent from the transfer processing section, encode the picture, and send to The transmission processing section transmits the coded data to specify the coded data. 31. The image transmission apparatus according to claim 30, wherein the transmission processing section includes a transmission buffer, and The transmission processing section receives the coded data specified by the image coding section at each multiplexing timing of the transmission buffer, and outputs the coded data in the order in which the coded data is received at a preset transmission rate , and the occupancy of encoded data in the transmission buffer is output as multiplex information. 32. The image transmission apparatus according to claim 30, wherein each image coding section comprises: a quantization width determination subsection; a quantized width image to encode the image, and The quantization width determining subsection calculates a buffer occupancy in each of a plurality of decoding devices that receive the transmitted from each image encoding section through the transmission processing section based on multiplexing information. encoding data, and setting the quantization width according to the calculated buffer occupancy in the decoding means and the number of encoded data not transmitted. 33. The image transmission device according to claim 30, wherein each image encoding section further includes an encoding data memory subsection for temporarily storing the encoded data, and The coded data memory subsection temporarily stores untransmitted data, which is data that is not transmitted to the transmission processing section at each transmission timing of coded data and uses as an upper limit by the image Each preset value of the encoding part is determined by subtracting the buffer occupancy in the decoding device calculated from the multiplexing information from the buffer size of the decoding device predicted from the receiving side. A value obtained transfers the coded data to the transfer processing section. 34. An image transmission device comprising a transmission processing section for receiving encoded data from a plurality of image encoding sections, Wherein the transmission processing section includes a transmission buffer, receives the coded data specified by the image coding section at each multiplexing timing of the transmission buffer, and receives the coded data at a preset transmission rate The coded data and the occupancy of the coded data in the transmission buffer are sequentially output as multiplexed information. 35. An image encoding apparatus comprising a quantization width determination section and a basic encoding processing section for encoding an image by quantizing the image with a quantization width specified by the quantization width determination section, Wherein the quantization width determination part is based on multiplexing information that can be used to calculate a delay number obtained until the coded data is transmitted to a virtual decoding device, the virtual decoding device calculated based on the multiplexing information The quantization width is determined by a buffer occupancy in , and a number of coded data not transmitted. 36. The image coding apparatus according to claim 35, further comprising a coded data memory section for temporarily storing the coded data, Wherein the encoded data memory section temporarily stores untransmitted data that is not transmitted per transmission timing of the encoded data, and uses a value preset by each image encoding section as an upper limit or passed from the receiver A value obtained by subtracting a buffer occupancy in the virtual decoding device calculated from the multiplexing information from a buffer size of the decoding device predicted by the side outputs the coded data. 37. An image transmission apparatus comprising a bit rate management section for managing a bit rate on a transmission line and one or more image encoding sections, wherein the bit rate management section determines the bit rate representing the sum of the transfer rates in one or more image coding sections on the transmission line to allow transmission of the code requested to be transmitted by any of the image coding sections at each transmission timing data, and transmits transmission information related to the transmission delay as a function of another picture coding portion allowed on the transmission line, and The one or more image encoding sections encode images by quantizing the images with a quantization width determined based on the transmission information, untransmitted encoded data, and a predicted buffer size on the receiving side. 38. An image transmission apparatus comprising a bit rate management section for managing the transmission rates of a plurality of image coding sections, wherein the bit rate management section determines the sum representing the transfer rates in all image coding sections on a transmission line, permits the transmission of coded data requested to be transmitted by each transfer timing of any image coding section, and transmits the Transmission information related to the transmission delay as a function of another picture coding portion allowed on the transmission line. 39. An image encoding method, comprising the steps of: (1) determining a quantization width; and (2) encoding the picture by quantizing the picture with the quantization width specified in step (1), Wherein in step (1) the quantization width is calculated based on the multiplexing information that can be used to calculate a delay number obtained until the coded data is transmitted to a virtual decoding device, based on the multiplexing information in the It is set according to a buffer occupancy in the virtual decoding device, and an amount of encoded data not transmitted. 40. The image encoding method according to claim 39, wherein untransmitted data which is not transmitted among the encoded data is temporarily stored at each transmission timing and used as an upper limit given by each part of the image encoding section. A value obtained by subtracting the buffer occupancy in the decoding device calculated from the multiplexing information from the buffer size of the decoding device predicted by the receiver side and outputting the coded data .

Images